
# RUN THIS CELL 
import requests
from IPython.core.display import HTML
styles = requests.get(
    "https://raw.githubusercontent.com/Harvard-IACS/2018-CS109A/master/"
    "content/styles/cs109.css"
).text
HTML(styles)


! cd $HOME
! git clone https://github.com/rafah1/COVID-data/

Cloning into 'COVID-data'...
remote: Enumerating objects: 63, done.
remote: Counting objects: 100% (5/5), done.
remote: Compressing objects: 100% (5/5), done.
remote: Total 63 (delta 0), reused 5 (delta 0), pack-reused 58
Unpacking objects: 100% (63/63), done.


pip install --upgrade gap-stat

Collecting gap-stat
  Downloading gap_stat-2.0.1-py3-none-any.whl (6.9 kB)
Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from gap-stat) (1.3.5)
Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from gap-stat) (1.4.1)
Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from gap-stat) (1.21.6)
Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas->gap-stat) (2.8.2)
Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->gap-stat) (2022.1)
Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas->gap-stat) (1.15.0)
Installing collected packages: gap-stat
Successfully installed gap-stat-2.0.1


# Import necessary libraries
import os
import time
import datetime
import requests
import glob
import numpy as np
import math
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import scipy.stats as stats
%matplotlib inline

# Tensorflow
import tensorflow as tf
from tensorflow.python.keras import backend as K

# Keras
from keras.preprocessing.sequence import TimeseriesGenerator

# sklearn
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from sklearn.metrics import mean_absolute_percentage_error, mean_squared_error, mean_absolute_error 
from sklearn.metrics import precision_score , recall_score
from sklearn.model_selection import train_test_split
from sklearn.model_selection import TimeSeriesSplit
from sklearn.metrics import classification_report
from sklearn.metrics import accuracy_score
from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import OneHotEncoder
from sklearn import manifold
from sklearn.impute import SimpleImputer, KNNImputer

#Cluster
import scipy.cluster.hierarchy as hac
from scipy.cluster.hierarchy import fcluster
from sklearn.cluster import KMeans, DBSCAN
from sklearn.preprocessing import StandardScaler
from sklearn.decomposition import PCA
from sklearn.metrics import silhouette_samples, silhouette_score
from sklearn.neighbors import NearestNeighbors
from gap_statistic import OptimalK

#Statistical models
from statsmodels.tsa.stattools import adfuller
from statsmodels.graphics.tsaplots import plot_acf
from statsmodels.tsa.arima_model import ARMA, ARIMA

#Date data
from datetime import datetime

#Numpy
from numpy.ma.core import reshape

#Table
from tabulate import tabulate

/usr/local/lib/python3.7/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.
  import pandas.util.testing as tm


# Verify setup
print("tensorflow version", tf.__version__)
print("keras version", tf.keras.__version__)
print("Eager Execution Enabled:", tf.executing_eagerly())

# Get the number of replicas 
strategy = tf.distribute.MirroredStrategy()
print("Number of replicas:", strategy.num_replicas_in_sync)

devices = tf.config.experimental.get_visible_devices()
print("Devices:", devices)
print(tf.config.experimental.list_logical_devices('GPU'))

print("GPU Available: ", tf.config.list_physical_devices('GPU'))
print("All Physical Devices", tf.config.list_physical_devices())

AUTOTUNE = tf.data.experimental.AUTOTUNE

tensorflow version 2.8.0
keras version 2.8.0
Eager Execution Enabled: True
WARNING:tensorflow:There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce.
INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',)
Number of replicas: 1
Devices: [PhysicalDevice(name='/physical_device:CPU:0', device_type='CPU')]
[]
GPU Available:  []
All Physical Devices [PhysicalDevice(name='/physical_device:CPU:0', device_type='CPU')]


#Check GPU
!nvidia-smi

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.


# Increase number of displayed columns
pd.set_option('display.max_columns', 700)
  
# Use glob to read the csv files from a folder
path = os.getcwd()
#csv_files = glob.glob(os.path.join("state_level_data", "*.csv"))
csv_files = glob.glob(os.path.join("./COVID-data/state_level_data", "*.csv"))
csv_files
  
# Read csv files
file_names=[]
list_files=[]
for f in csv_files:      
    # Print name of file
    name=f.split("/")[-1]
    name=name.split(".")[0]
    #print(f'\033[1mFILE NAME: {name} \033[0m')  
    file_names.append(name)
    # read the csv files
    df = pd.read_csv(f)
    list_files.append(df)
    ####
    ## Delete the "#" below to print shape, description, data type and NAs of each individual csv file
    ####
    #print(f'\nShape: {df.shape}')
    #print('\nContent:')
    #display(df.head())
    #print('\nDescription:')
    #display(df.describe())
    #print('\nData type:')
    #display(np.transpose(pd.DataFrame(df.dtypes)))
    #print("\nNumber of NAs per column:")
    #display(np.transpose(pd.DataFrame(df.isnull().sum())))
    #print("\n")

#Display one state
print("The shape of the file of one state is", list_files[0].shape)
print("\nThe first rows of the file of one state are:")
display(list_files[0].head())

The shape of the file of one state is (744, 499)

The first rows of the file of one state are:


# Include state name as column in each csv file
n=0
for i in list_files:
    i["State_Name"]=file_names[n]
    n=n+1

# Print file before cleaning
print("\nNumber of NAs per column of the file of one State before cleaning:")
display(np.transpose(pd.DataFrame(list_files[0].isnull().sum())))

list_files_clean=[]
#Clean data 
for i in list_files:
    i.interpolate('index', limit_direction='both', inplace=True)
    list_files_clean.append(i) 

# Print file after cleaning
print("\nNumber of NAs per column of the file of one State after cleaning:")
display(np.transpose(pd.DataFrame(list_files_clean[0].isnull().sum())))

Number of NAs per column of the file of one State before cleaning:

Number of NAs per column of the file of one State after cleaning:


# Join all files
df_all=list_files_clean[0]
for i in range(1,len(list_files_clean)):
    df_all=pd.concat([df_all, list_files_clean[i]], axis=0)
df_all=df_all.fillna(0) 

# Convert date column from object to date format
df_all['date'] = pd.to_datetime(df_all['date'], dayfirst=True)

# Create one hot encoder for state    
df_f=pd.get_dummies(df_all, columns=["State_Name"]) 

# Print merged file
print(f'\nShape of merged file with 50 States: {df_f.shape}')
print('\nContent of merged file with 50 States:')
display(df_f.head())
print('\nDescription of merged file with 50 States:')
display(df_f.describe())
print('\nData type of merged file with 50 States:')
display(np.transpose(pd.DataFrame(df_f.dtypes)))
print("\nNumber of NAs per column of merged file with 50 States after cleaning:")
display(np.transpose(pd.DataFrame(df_f.isnull().sum())))

Shape of merged file with 50 States: (37201, 550)

Content of merged file with 50 States:

Description of merged file with 50 States:

Data type of merged file with 50 States:

Number of NAs per column of merged file with 50 States after cleaning:


##General variables
# List variables
var_inc=list(df_f.columns)
del var_inc[0] #delete date as variable

# List state names
state_names=["State_Name_"+s for s in file_names]

# Sort file names
#file_names.sort()


##Function for non-normalized dataframe of the selected state

def sel_state(state):    
    # Index of selected state
    sel_state=file_names.index(state)

    # Dataset selected state
    df_nnormal=df_f.copy()
    df_state=df_nnormal[df_nnormal[state_names[sel_state]]==1]
     
    return sel_state, df_state

## Run function for getting datasets of the selected state
info_sel_state = sel_state("New York")

## Print shape of file
print("Shape of non-normalized file of selected state:", info_sel_state[1].shape)

Shape of non-normalized file of selected state: (744, 550)


def train_test(state):    
    # Dataset selected state
    info_sel_state=sel_state(state)
    df_state=info_sel_state[1]
    

    # Exclude state name and date columns
    df_state.drop(state_names, axis=1, inplace=True)
    df_date=df_state["date"]
    df_state.drop("date", axis=1, inplace=True)
    col_names=list(df_state.columns)
    del col_names[0]

    #Exclude y
    df_y=df_state[["JHU_cases"]]
    df_state.drop("JHU_cases", axis=1, inplace=True)

    #Create train data (75%) and test data (25%)
    data_75=len(df_state)*3//4
    xtrain=df_state[:data_75]
    ytrain=df_y[:data_75]

    xtest=df_state[data_75:]
    ytest=df_y[data_75:]

    # Normalize xtrain dataframe
    scaler_x = MinMaxScaler(feature_range=(0, 1))
    scaled_xtrain = scaler_x.fit_transform(xtrain)
    scaled_df_xtrain = pd.DataFrame(scaled_xtrain)
    scaled_df_xtrain.columns=col_names

    #Normalized xtest dataframe
    scaled_xtest=scaler_x.transform(xtest)
    scaled_df_xtest = pd.DataFrame(scaled_xtest)
    scaled_df_xtest.columns=col_names

    # Add date column to normalized datasets
    df_date_xtrain=df_date[:data_75]
    df_normal_xtrain=pd.merge(scaled_df_xtrain.reset_index(drop=True), df_date_xtrain.reset_index(drop=True),left_index=True, right_index=True)
    
    df_date_xtest=df_date[data_75:]
    df_normal_xtest=pd.merge(scaled_df_xtest.reset_index(drop=True), df_date_xtest.reset_index(drop=True),left_index=True, right_index=True)

    # Add y column to normalized datasets
    df_normal_xtrain=pd.merge(ytrain.reset_index(drop=True),df_normal_xtrain.reset_index(drop=True), left_index=True, right_index=True)
    df_normal_xtrain['date'] = pd.to_datetime(df_normal_xtrain['date'], dayfirst=True)

    df_normal_xtest=pd.merge(ytest.reset_index(drop=True),df_normal_xtest.reset_index(drop=True), left_index=True, right_index=True)
    df_normal_xtest['date'] = pd.to_datetime(df_normal_xtest['date'], dayfirst=True)

    #Reshape x data
    x_train=np.array(scaled_df_xtrain)
    x_train = np.reshape(x_train, (x_train.shape[0], x_train.shape[1], 1))

    x_test=np.array(scaled_df_xtest)
    x_test = np.reshape(x_test, (x_test.shape[0], x_test.shape[1], 1))

    return x_train, ytrain["JHU_cases"], x_test, ytest["JHU_cases"], df_normal_xtrain, df_normal_xtest

## Run function for getting datasets of the selected state
info_train_test = train_test("New York")

## Print shape of files
print("Example of dataset created for one selected state (New York):\n")
print("Shape of x_train of selected state:", info_train_test[0].shape)
print("Shape of y_train of selected state:", info_train_test[1].shape)
print("Shape of x_test of selected state:", info_train_test[2].shape)
print("Shape of y_test of selected state:", info_train_test[3].shape)

Example of dataset created for one selected state (New York):

Shape of x_train of selected state: (558, 498, 1)
Shape of y_train of selected state: (558,)
Shape of x_test of selected state: (186, 498, 1)
Shape of y_test of selected state: (186,)


#Selected state NY and call variables
info_sel_state = sel_state("New York")
info_train_test = train_test("New York")
ny_df=info_train_test[4] # This is a DF of xtrain with the information pertaining to only the state of NY
ytrain=info_train_test[1]

#Simple moving average for JHU Cases (index 0) for a size of 3 and create a new column to store it
ny_df['pandas_SMA_3'] = ny_df.iloc[:,0].rolling(window=3).mean() #Note 0 is the column for JHU_cases    

#Simple moving average for a size of 20 and create a new column to store it
#Using the clean implementation
ny_df['pandas_SMA_20'] = ny_df.iloc[:,0].rolling(window=20).mean()

#Doing the plots
plt.figure(figsize=(12,6))
plt.title("Simple Moving Averages to predict JHU Covid Cases in New York")
plt.plot(ny_df['date'],ny_df['JHU_cases'],label='Covid Cases')
plt.plot(ny_df['date'],ny_df['pandas_SMA_3'],label='SMA 3 Days')
plt.plot(ny_df['date'],ny_df['pandas_SMA_20'],label='SMA 20 Days')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


# Cumulative Moving Average for JHU Cases (index 0) for 4 min periods and create a new column to store it
ny_df['CMA_4'] = ny_df['JHU_cases'].expanding(min_periods=4).mean()

plt.figure(figsize=(12,6))
plt.title("Cumulative Moving Average to predict JHU Covid Cases in New York")
plt.plot(ny_df['date'],ny_df['JHU_cases'],label='Covid Cases')
plt.plot(ny_df['date'],ny_df['CMA_4'],label='CMA with 4 periods')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


ny_df['EMA'] = ny_df.iloc[:,0].ewm(span=40,adjust=False).mean()

plt.figure(figsize=(12,6))
plt.title("Exponential Moving Average to predict JHU Covid Cases in New York")
plt.plot(ny_df['date'],ny_df['JHU_cases'],label='Covid Cases')
plt.plot(ny_df['date'],ny_df['EMA'],label='EMA')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


def return_metrics(input_df):
    input_df = input_df.fillna(0) #replace nan that appeared in the first poisitions of SMA
    
    MAPE_SMA3=mean_absolute_error(input_df['JHU_cases'].tolist(), input_df['pandas_SMA_3'].tolist())
    MAPE_SMA20=mean_absolute_error(input_df['JHU_cases'].tolist(), input_df['pandas_SMA_20'].tolist())
    MAPE_CMA4=mean_absolute_error(input_df['JHU_cases'].tolist(), input_df['CMA_4'].tolist())
    MAPE_EMA=mean_absolute_error(input_df['JHU_cases'].tolist(), input_df['EMA'].tolist())
    print("-------------------------------------")
    print("Mean Absolute Error (MAE)")
    print("For Simple Moving Average (SMA) size of 3:",MAPE_SMA3)
    print("For Simple Moving Average (SMA) size of 20:",MAPE_SMA20)
    print("For Cumulative Moving Average (CMA) with 4 minimum periods:",MAPE_CMA4)
    print("For Exponential Moving Average (EMA):",MAPE_EMA)
        
    MSE_SMA3=mean_squared_error(input_df['JHU_cases'].tolist(), input_df['pandas_SMA_3'].tolist())
    MSE_SMA20=mean_squared_error(input_df['JHU_cases'].tolist(), input_df['pandas_SMA_20'].tolist())
    MSE_CMA4=mean_squared_error(input_df['JHU_cases'].tolist(), input_df['CMA_4'].tolist())
    MSE_EMA=mean_squared_error(input_df['JHU_cases'].tolist(), input_df['EMA'].tolist())
    print("-------------------------------------")
    print("Mean Squared Error (MSE)")
    print("For Simple Moving Average (SMA) size of 3:",MSE_SMA3)
    print("For Simple Moving Average (SMA) size of 20:",MSE_SMA20)
    print("For Cumulative Moving Average (CMA) with 4 minimum periods:",MSE_CMA4)
    print("For Exponential Moving Average (EMA):",MSE_EMA)
    print("-------------------------------------")
        
    return MAPE_SMA3, MAPE_SMA20, MAPE_CMA4, MAPE_EMA

print("Calculating performance of Moving Averages for Covid Cases in New York")
metrics_ny=return_metrics(ny_df)
# Note that according to https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_absolute_percentage_error.html
# the value of MAPE when some element of the y_true is zero is arbitrarily high because
# of the division by epsilon

Calculating performance of Moving Averages for Covid Cases in New York
-------------------------------------
Mean Absolute Error (MAE)
For Simple Moving Average (SMA) size of 3: 456.59259259259255
For Simple Moving Average (SMA) size of 20: 1035.6623655913977
For Cumulative Moving Average (CMA) with 4 minimum periods: 2907.3488355640297
For Exponential Moving Average (EMA): 1430.9678843660217
-------------------------------------
Mean Squared Error (MSE)
For Simple Moving Average (SMA) size of 3: 1126065.548387097
For Simple Moving Average (SMA) size of 20: 3538284.6018458786
For Cumulative Moving Average (CMA) with 4 minimum periods: 19145351.075307716
For Exponential Moving Average (EMA): 5111197.247293777
-------------------------------------


# Augmented Dickey Fuller test (ADF Test) is a common statistical test 
# used to test whether a given Time series is stationary or not. 
# Please refer to reference (5) in the references section for further detail

# Using info New York
info_sel_state = sel_state("New York")
info_train_test=train_test("New York")

# scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# normalizing ytrain
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytrain = ytrain.reshape(1,-1)
ytrain = pd.Series(ytrain[0])


ADF = adfuller(ytrain)   
print(f'ADF Statistic: {round(ADF[0],3)}') 
print(f'p-value: {ADF[1]:.3e}')

ADF Statistic: -2.414
p-value: 1.379e-01


# Differentiating ytrain to the first order by using np.diff function
ytrain_diff = np.diff(ytrain, n=1)

ADF_diff = adfuller(ytrain_diff)   
print(f'ADF Statistic: {round(ADF_diff[0],3)}') 
print(f'p-value: {ADF_diff[1]:.3e}')

ADF Statistic: -6.031
p-value: 1.418e-07


plot_acf(ytrain_diff, lags=40);
plt.tight_layout()
plt.xlabel("Lags")
plt.ylabel("Autocorrelation")

Text(18.75, 0.5, 'Autocorrelation')


#############
# ARMA VALIDATION PART
#############
# Please refer to references (6) and (7) in the reference section
# for further details regarding the code used in this section


###
#xtrain
###
xtrain=info_train_test[4]
xtrain_copy=xtrain.loc[:, ~xtrain.columns.isin(['date', 'JHU_cases'])]

model_ma_results=[]
models_ma=[]
rmse=[]
# we are specifying the hyper parameters of the MA model below
# we will use 1st, 3rd and 7th lag in the model
rmse_list = [1,3,7]

# Time series cross-validation 

tss = TimeSeriesSplit(n_splits = 4, gap=0)
# loop for hyperparameters
for i in rmse_list:
  # loop for cross validation
  for train_index, val_index in tss.split(ytrain):
      train_y, val_y = ytrain.iloc[train_index], ytrain.iloc[val_index]
      
      # we will also be using the first order AR in this model 
      model_ma = ARMA(train_y, order=(1, i)).fit(disp=False)
      predictions = model_ma.predict(val_y.index.values[0], val_y.index.values[-1])
      true_values = val_y.values
      # appending the rmse to the list
      # we will pick the lag with the smallest rmse 
      rmse.append(math.sqrt(mean_squared_error(true_values, predictions)))

  print(f"RMSE for order {i} is : ", round(np.mean(rmse),4))

RMSE for order 1 is :  0.1367
RMSE for order 3 is :  0.1331
RMSE for order 7 is :  0.1325


# ARMA TEST PART

# normalizing ytest
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)
ytest = ytest.reshape(1,-1)
ytest = pd.Series(ytest[0])

# using the 1st order AR and 7th order MA for test
model_ma = ARMA(train_y, order=(1, 7)).fit(disp=False)

ytest = ytest.reset_index(drop=True)
predictions = model_ma.predict(ytest.index.values[0], ytest.index.values[-1])

true_values = ytest.values
print("The MSE is : " , round(math.sqrt(mean_squared_error(true_values, predictions)),4))
MAE_0_ny=mean_absolute_error(true_values, predictions)
print("The MAE is : " , round(MAE_0_ny,4))

The MSE is :  0.2144
The MAE is :  0.1295


# The predictions below are the predictions for order 7
# We reshape the predictions then 
# inverse transform and plot the predictions vs. actuals 
predictions = predictions.values
predictions = predictions.reshape(-1, 1)
list_ARMA = scaler_y.inverse_transform(predictions).reshape(1,-1)[0]
#list_ARMA = predictions.tolist()

ytest  = info_train_test[3]

#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)

plt.figure(figsize=(12,6))
plt.plot(range(len(ytest)), list_ARMA, label="Predicted")
plt.plot(range(len(ytest)),list(ytest), label="Actual")
plt.title(f"AR (1) MA (7) Predictions JHU Covid Cases for {file_names[info_sel_state[0]]}")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Selected state California and call variables
info_sel_state = sel_state("California")
info_train_test = train_test("California")
ca_df=info_train_test[4] # This is a DF of xtrain with the information pertaining to only the state of NY
ytrain=info_train_test[1]

#Simple moving average for JHU Cases (index 0) for a size of 3 and create a new column to store it
ca_df['pandas_SMA_3'] = ca_df.iloc[:,0].rolling(window=3).mean() #Note 0 is the column for JHU_cases    

#Simple moving average for a size of 20 and create a new column to store it
#Using the clean implementation
ca_df['pandas_SMA_20'] = ca_df.iloc[:,0].rolling(window=20).mean()

#Doing the plots
plt.figure(figsize=(12,6))
plt.title("Simple Moving Averages to predict JHU Covid Cases in California")
plt.plot(ca_df['date'],ca_df['JHU_cases'],label='Covid Cases')
plt.plot(ca_df['date'],ca_df['pandas_SMA_3'],label='SMA 3 Days')
plt.plot(ca_df['date'],ca_df['pandas_SMA_20'],label='SMA 20 Days')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


# Cumulative Moving Average for JHU Cases (index 0) for 4 min periods and create a new column to store it
ca_df['CMA_4'] = ca_df['JHU_cases'].expanding(min_periods=4).mean()

plt.figure(figsize=(12,6))
plt.title("Cumulative Moving Average to predict JHU Covid Cases in California")
plt.plot(ca_df['date'],ca_df['JHU_cases'],label='Covid Cases')
plt.plot(ca_df['date'],ca_df['CMA_4'],label='CMA with 4 periods')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


ca_df['EMA'] = ca_df.iloc[:,0].ewm(span=40,adjust=False).mean()

plt.figure(figsize=(12,6))
plt.title("Exponential Moving Average to predict JHU Covid Cases in California")
plt.plot(ca_df['date'],ca_df['JHU_cases'],label='Covid Cases')
plt.plot(ca_df['date'],ca_df['EMA'],label='EMA')
plt.xlabel("Date")
plt.ylabel("JHU cases")
plt.legend(loc=2)
plt.show()


print("Calculating performance of Moving Averages for Covid Cases in California")
metrics_cal=return_metrics(ca_df)

# Note that according to https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_absolute_percentage_error.html
# the value of MAPE when some element of the y_true is zero is arbitrarily high because
# of the division by epsilon

Calculating performance of Moving Averages for Covid Cases in California
-------------------------------------
Mean Absolute Error (MAE)
For Simple Moving Average (SMA) size of 3: 920.7909199522103
For Simple Moving Average (SMA) size of 20: 2080.134677419355
For Cumulative Moving Average (CMA) with 4 minimum periods: 6010.752394787415
For Exponential Moving Average (EMA): 3058.7360460414293
-------------------------------------
Mean Squared Error (MSE)
For Simple Moving Average (SMA) size of 3: 4123623.7256073286
For Simple Moving Average (SMA) size of 20: 17972695.588239245
For Cumulative Moving Average (CMA) with 4 minimum periods: 117571636.7037175
For Exponential Moving Average (EMA): 31312807.78724574
-------------------------------------


# Augmented Dickey Fuller test (ADF Test) is a common statistical test 
# used to test whether a given Time series is stationary or not. 
# Please refer to reference (5) in the references section for further detail

# ytrain

# Using info New York
info_sel_state = sel_state("California")
info_train_test=train_test("California")

#ytrain=info_train_test[1]

# scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# normalizing ytrain
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytrain = ytrain.reshape(1,-1)
ytrain = pd.Series(ytrain[0])

ADF = adfuller(ytrain)   
print(f'ADF Statistic: {round(ADF[0],3)}') 
print(f'p-value: {ADF[1]:.3e}')

ADF Statistic: -2.732
p-value: 6.863e-02


# Differentiating ytrain to the first order by using np.diff function
ytrain_diff = np.diff(ytrain, n=1)

ADF_diff = adfuller(ytrain_diff)   
print(f'ADF Statistic: {round(ADF_diff[0],3)}') 
print(f'p-value: {ADF_diff[1]:.3e}')

ADF Statistic: -4.343
p-value: 3.741e-04


from statsmodels.graphics.tsaplots import plot_acf
plot_acf(ytrain_diff, lags=40);
plt.tight_layout()
plt.xlabel("Lags")
plt.ylabel("Autocorrelation")

Text(18.75, 0.5, 'Autocorrelation')


#############
# ARMA VALIDATION PART
#############
# Please refer to references (6) and (7) in the reference section
# for further details regarding the code used in this section


###
#xtrain
###
xtrain=info_train_test[4]
xtrain_copy=xtrain.loc[:, ~xtrain.columns.isin(['date', 'JHU_cases'])]

from statsmodels.tsa.arima_model import ARMA
from sklearn.metrics import mean_squared_error
import math

model_ma_results=[]
models_ma=[]
rmse=[]
# we are specifying the hyper parameters of the MA model below
# we will use 1st, 3rd and 7th lag in the model
rmse_list = [1,4,5,6,7]

# Time series cross-validation 

tss = TimeSeriesSplit(n_splits = 3, gap=0)
# loop for hyperparameters
for i in rmse_list:
  # loop for cross validation
  for train_index, val_index in tss.split(ytrain):
      train_y, val_y = ytrain.iloc[train_index], ytrain.iloc[val_index]
      
      # we will be using AR(0) in this model 
      model_ma = ARMA(train_y, order=(0, i)).fit(disp=False)
      predictions = model_ma.predict(val_y.index.values[0], val_y.index.values[-1])
      true_values = val_y.values
      # appending the rmse to the list
      # we will pick the lag with the smallest rmse 
      rmse.append(math.sqrt(mean_squared_error(true_values, predictions)))

  print(f"RMSE for order {i} is : ", round(np.mean(rmse),4))

RMSE for order 1 is :  0.188
RMSE for order 4 is :  0.1879
RMSE for order 5 is :  0.1878
RMSE for order 6 is :  0.1878
RMSE for order 7 is :  0.1877


# ARMA TEST PART

# normalizing ytest
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)
ytest = ytest.reshape(1,-1)
ytest = pd.Series(ytest[0])

# using the 0th order AR and 7th order MA for test
model_ma = ARMA(train_y, order=(0, 7)).fit(disp=False)

ytest = ytest.reset_index(drop=True)
predictions = model_ma.predict(ytest.index.values[0], ytest.index.values[-1])

true_values = ytest.values
print("The MSE is : " , round(math.sqrt(mean_squared_error(true_values, predictions)),4))
MAE_0_cal=mean_absolute_error(true_values, predictions)
print("The MAE is : " , round(MAE_0_cal,4))

The MSE is :  0.1111
The MAE is :  0.0658


# The predictions below are the predictions for order 7
# We reshape the predictions then 
# inverse transform and plot the predictions vs. actuals 
predictions = predictions.values
predictions = predictions.reshape(-1, 1)
list_ARMA = scaler_y.inverse_transform(predictions).reshape(1,-1)[0]
#list_ARMA = predictions.tolist()

ytest  = info_train_test[3]

#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)

plt.figure(figsize=(12,6))
plt.plot(range(len(ytest)), list_ARMA, label="Predicted")
plt.plot(range(len(ytest)),list(ytest), label="Actual")
plt.title(f"AR (0) MA (7) Predictions JHU Covid Cases for {file_names[info_sel_state[0]]}")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


tf.keras.backend.clear_session()

# Using info New York
info_sel_state = sel_state("New York")
info_train_test=train_test("New York")

# Scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# minmax scaler for y, scaling between 0 and 1
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)

#Call x variables
xtrain=info_train_test[4]
xtrain.drop(["date", "JHU_cases"], axis=1, inplace=True)
xtest=info_train_test[5].copy()
xtest.drop(["date", "JHU_cases"], axis=1, inplace=True)

# time series generator uses 14 day block to predict the next day 
train_gen = TimeseriesGenerator(xtrain.to_numpy(), ytrain,
                               length=14, sampling_rate=1,  
                                batch_size = 558)

test_gen = TimeseriesGenerator(xtest.to_numpy(), ytest,
                                length=14, sampling_rate=1, 
                               batch_size = 186)

# Below we are creating x and y train and test variable from the generators
x_train, y_train = train_gen[0]
# x_val, y_val = val_gen[0]
x_test, y_test = test_gen[0]

print('x train shape is :', x_train.shape)
print('y train shape is :', y_train.shape)
print('x test shape is :' , x_test.shape)
print('y test shape is :' , y_test.shape)

x train shape is : (544, 14, 498)
y train shape is : (544, 1)
x test shape is : (172, 14, 498)
y test shape is : (172, 1)


# input dimension below
input_dim = x_train.shape[1:]
n_units = 100

#Create model
model_rnn_input = tf.keras.Input(shape=input_dim)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units, return_sequences=False)(model_rnn_input)
model_rnn_output = tf.keras.layers.Dense(units=1, activation='linear')(model_rnn_hidden)
model_rnn = tf.keras.Model(inputs=model_rnn_input, outputs=model_rnn_output, name="model_rnn")

#Print the model architecture
print(model_rnn.summary())

#Compile model
model_rnn.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_rnn.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_rnn"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 simple_rnn (SimpleRNN)      (None, 100)               59900     
                                                                 
 dense (Dense)               (None, 1)                 101       
                                                                 
=================================================================
Total params: 60,001
Trainable params: 60,001
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 4s 5ms/step - loss: 0.0375 - mae: 0.1197 - val_loss: 0.0363 - val_mae: 0.1828
Epoch 2/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0071 - mae: 0.0597 - val_loss: 0.0198 - val_mae: 0.1352
Epoch 3/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0070 - mae: 0.0598 - val_loss: 0.0097 - val_mae: 0.0935
Epoch 4/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0053 - mae: 0.0484 - val_loss: 0.0155 - val_mae: 0.1207
Epoch 5/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0038 - mae: 0.0426 - val_loss: 0.0185 - val_mae: 0.1322
Epoch 6/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0045 - mae: 0.0480 - val_loss: 0.0104 - val_mae: 0.0971
Epoch 7/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0040 - mae: 0.0443 - val_loss: 0.0117 - val_mae: 0.1046
Epoch 8/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0050 - mae: 0.0499 - val_loss: 0.0018 - val_mae: 0.0352
Epoch 9/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0036 - mae: 0.0423 - val_loss: 0.0019 - val_mae: 0.0366
Epoch 10/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0035 - mae: 0.0435 - val_loss: 0.0199 - val_mae: 0.1389
Epoch 11/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0053 - mae: 0.0491 - val_loss: 0.0018 - val_mae: 0.0378
Epoch 12/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0038 - mae: 0.0442 - val_loss: 0.0050 - val_mae: 0.0681
Epoch 13/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0034 - mae: 0.0449 - val_loss: 9.0919e-04 - val_mae: 0.0241
Epoch 14/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0040 - mae: 0.0457 - val_loss: 0.0097 - val_mae: 0.0945
Epoch 15/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0037 - mae: 0.0435 - val_loss: 8.1231e-04 - val_mae: 0.0241
Epoch 16/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0024 - mae: 0.0344 - val_loss: 0.0048 - val_mae: 0.0648
Epoch 17/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0047 - mae: 0.0522 - val_loss: 0.0063 - val_mae: 0.0755
Epoch 18/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0023 - mae: 0.0327 - val_loss: 0.0011 - val_mae: 0.0262
Epoch 19/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0027 - mae: 0.0386 - val_loss: 0.0019 - val_mae: 0.0400
Epoch 20/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0024 - mae: 0.0356 - val_loss: 0.0015 - val_mae: 0.0337
Epoch 21/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0021 - mae: 0.0333 - val_loss: 9.5373e-04 - val_mae: 0.0254
Epoch 22/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0027 - mae: 0.0390 - val_loss: 8.8964e-04 - val_mae: 0.0239
Epoch 23/50
489/489 [==============================] - 3s 7ms/step - loss: 0.0019 - mae: 0.0320 - val_loss: 5.5364e-04 - val_mae: 0.0193
Epoch 24/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0030 - mae: 0.0407 - val_loss: 0.0043 - val_mae: 0.0634
Epoch 25/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0023 - mae: 0.0363 - val_loss: 6.7585e-04 - val_mae: 0.0204
Epoch 26/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0033 - mae: 0.0430 - val_loss: 0.0077 - val_mae: 0.0850
Epoch 27/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0028 - mae: 0.0374 - val_loss: 0.0112 - val_mae: 0.1024
Epoch 28/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0038 - mae: 0.0457 - val_loss: 0.0012 - val_mae: 0.0267
Epoch 29/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0034 - mae: 0.0393 - val_loss: 0.0186 - val_mae: 0.1323
Epoch 30/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0029 - mae: 0.0377 - val_loss: 0.0032 - val_mae: 0.0463
Epoch 31/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0033 - mae: 0.0399 - val_loss: 7.6156e-04 - val_mae: 0.0210
Epoch 32/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0015 - mae: 0.0278 - val_loss: 0.0037 - val_mae: 0.0555
Epoch 33/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0020 - mae: 0.0325 - val_loss: 0.0025 - val_mae: 0.0424
Epoch 34/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0020 - mae: 0.0334 - val_loss: 0.0012 - val_mae: 0.0261
Epoch 35/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0025 - mae: 0.0381 - val_loss: 0.0014 - val_mae: 0.0345
Epoch 36/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0027 - mae: 0.0379 - val_loss: 3.4524e-04 - val_mae: 0.0144
Epoch 37/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0018 - mae: 0.0307 - val_loss: 0.0016 - val_mae: 0.0342
Epoch 38/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0017 - mae: 0.0282 - val_loss: 0.0014 - val_mae: 0.0344
Epoch 39/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0033 - mae: 0.0404 - val_loss: 0.0156 - val_mae: 0.1237
Epoch 40/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0017 - mae: 0.0308 - val_loss: 5.9440e-04 - val_mae: 0.0201
Epoch 41/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0019 - mae: 0.0316 - val_loss: 0.0087 - val_mae: 0.0925
Epoch 42/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0021 - mae: 0.0344 - val_loss: 0.0032 - val_mae: 0.0478
Epoch 43/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0012 - mae: 0.0265 - val_loss: 0.0015 - val_mae: 0.0325
Epoch 44/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0031 - mae: 0.0404 - val_loss: 0.0043 - val_mae: 0.0576
Epoch 45/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0023 - mae: 0.0356 - val_loss: 4.1848e-04 - val_mae: 0.0153
Epoch 46/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0016 - mae: 0.0283 - val_loss: 0.0087 - val_mae: 0.0860
Epoch 47/50
489/489 [==============================] - 3s 6ms/step - loss: 0.0024 - mae: 0.0375 - val_loss: 0.0014 - val_mae: 0.0356
Epoch 48/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0024 - mae: 0.0369 - val_loss: 8.4325e-04 - val_mae: 0.0257
Epoch 49/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0018 - mae: 0.0307 - val_loss: 0.0056 - val_mae: 0.0734
Epoch 50/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0027 - mae: 0.0351 - val_loss: 0.0012 - val_mae: 0.0271


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of one-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of one-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
rnn_prediction = scaler_y.inverse_transform(model_rnn.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), rnn_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (one-layer RNN model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_1=round(mean_MAE[0],4)
MAE_val_1=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_1)
print("The validation MAE of the model is", MAE_val_1)

The MAE of the model is 0.0408
The validation MAE of the model is 0.0602


# input dimension below
input_dim = x_train.shape[1:]
n_units = 128
n_units2 = 64

#Create model
model_rnn_input = tf.keras.Input(shape=input_dim)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units, return_sequences=True, dropout=0.2 )(model_rnn_input)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units2, return_sequences=False, dropout=0.2)(model_rnn_hidden)
model_rnn_output = tf.keras.layers.Dense(units=1, activation='linear')(model_rnn_hidden)
model_rnn = tf.keras.Model(inputs=model_rnn_input, outputs=model_rnn_output, name="model_rnn")


#Print the model architecture
print(model_rnn.summary())

#Compile model
model_rnn.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_rnn.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_rnn"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 simple_rnn_1 (SimpleRNN)    (None, 14, 128)           80256     
                                                                 
 simple_rnn_2 (SimpleRNN)    (None, 64)                12352     
                                                                 
 dense_1 (Dense)             (None, 1)                 65        
                                                                 
=================================================================
Total params: 92,673
Trainable params: 92,673
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 6s 9ms/step - loss: 0.0792 - mae: 0.1850 - val_loss: 0.1080 - val_mae: 0.3281
Epoch 2/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0166 - mae: 0.0988 - val_loss: 0.0021 - val_mae: 0.0421
Epoch 3/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0100 - mae: 0.0734 - val_loss: 0.0062 - val_mae: 0.0768
Epoch 4/50
489/489 [==============================] - 5s 10ms/step - loss: 0.0103 - mae: 0.0730 - val_loss: 0.0494 - val_mae: 0.2205
Epoch 5/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0081 - mae: 0.0649 - val_loss: 0.0239 - val_mae: 0.1532
Epoch 6/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0091 - mae: 0.0678 - val_loss: 0.0047 - val_mae: 0.0670
Epoch 7/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0072 - mae: 0.0590 - val_loss: 8.1806e-04 - val_mae: 0.0247
Epoch 8/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0077 - mae: 0.0638 - val_loss: 0.0134 - val_mae: 0.1129
Epoch 9/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0079 - mae: 0.0601 - val_loss: 0.0292 - val_mae: 0.1700
Epoch 10/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0076 - mae: 0.0623 - val_loss: 0.0045 - val_mae: 0.0576
Epoch 11/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0060 - mae: 0.0534 - val_loss: 0.0175 - val_mae: 0.1221
Epoch 12/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0070 - mae: 0.0587 - val_loss: 0.0068 - val_mae: 0.0813
Epoch 13/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0062 - mae: 0.0584 - val_loss: 2.2526e-04 - val_mae: 0.0096
Epoch 14/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0091 - mae: 0.0657 - val_loss: 5.1515e-04 - val_mae: 0.0193
Epoch 15/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0084 - mae: 0.0658 - val_loss: 0.0035 - val_mae: 0.0556
Epoch 16/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0076 - mae: 0.0598 - val_loss: 0.0081 - val_mae: 0.0848
Epoch 17/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0060 - mae: 0.0520 - val_loss: 0.0073 - val_mae: 0.0842
Epoch 18/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0064 - mae: 0.0568 - val_loss: 5.8328e-04 - val_mae: 0.0223
Epoch 19/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0073 - mae: 0.0611 - val_loss: 0.0086 - val_mae: 0.0844
Epoch 20/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0068 - mae: 0.0575 - val_loss: 0.0072 - val_mae: 0.0811
Epoch 21/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0073 - mae: 0.0588 - val_loss: 0.0112 - val_mae: 0.0989
Epoch 22/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0068 - mae: 0.0592 - val_loss: 0.0113 - val_mae: 0.1016
Epoch 23/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0058 - mae: 0.0532 - val_loss: 0.0017 - val_mae: 0.0359
Epoch 24/50
489/489 [==============================] - 4s 9ms/step - loss: 0.0066 - mae: 0.0584 - val_loss: 0.0031 - val_mae: 0.0445
Epoch 25/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0053 - mae: 0.0502 - val_loss: 0.0127 - val_mae: 0.1117
Epoch 26/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0058 - mae: 0.0550 - val_loss: 0.0188 - val_mae: 0.1363
Epoch 27/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0072 - mae: 0.0600 - val_loss: 0.0114 - val_mae: 0.1016
Epoch 28/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0071 - mae: 0.0598 - val_loss: 0.0035 - val_mae: 0.0478
Epoch 29/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0064 - mae: 0.0557 - val_loss: 8.8842e-04 - val_mae: 0.0244
Epoch 30/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0052 - mae: 0.0502 - val_loss: 0.0082 - val_mae: 0.0889
Epoch 31/50
489/489 [==============================] - 5s 10ms/step - loss: 0.0074 - mae: 0.0605 - val_loss: 0.0064 - val_mae: 0.0723
Epoch 32/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0064 - mae: 0.0530 - val_loss: 0.0021 - val_mae: 0.0402
Epoch 33/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0047 - mae: 0.0458 - val_loss: 0.0022 - val_mae: 0.0431
Epoch 34/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0052 - mae: 0.0497 - val_loss: 0.0027 - val_mae: 0.0472
Epoch 35/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0056 - mae: 0.0525 - val_loss: 0.0118 - val_mae: 0.1080
Epoch 36/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0058 - mae: 0.0557 - val_loss: 0.0036 - val_mae: 0.0545
Epoch 37/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0056 - mae: 0.0503 - val_loss: 0.0073 - val_mae: 0.0801
Epoch 38/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0055 - mae: 0.0527 - val_loss: 0.0020 - val_mae: 0.0428
Epoch 39/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0048 - mae: 0.0485 - val_loss: 0.0105 - val_mae: 0.1017
Epoch 40/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0052 - mae: 0.0478 - val_loss: 0.0016 - val_mae: 0.0360
Epoch 41/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0056 - mae: 0.0518 - val_loss: 0.0080 - val_mae: 0.0875
Epoch 42/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0050 - mae: 0.0484 - val_loss: 0.0361 - val_mae: 0.1644
Epoch 43/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0077 - mae: 0.0603 - val_loss: 2.7831e-04 - val_mae: 0.0140
Epoch 44/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0047 - mae: 0.0462 - val_loss: 0.0087 - val_mae: 0.0879
Epoch 45/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0066 - mae: 0.0558 - val_loss: 0.0159 - val_mae: 0.1234
Epoch 46/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0054 - mae: 0.0493 - val_loss: 5.1348e-04 - val_mae: 0.0191
Epoch 47/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0049 - mae: 0.0482 - val_loss: 0.0012 - val_mae: 0.0284
Epoch 48/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0055 - mae: 0.0522 - val_loss: 0.0025 - val_mae: 0.0464
Epoch 49/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0068 - mae: 0.0593 - val_loss: 0.0308 - val_mae: 0.1746
Epoch 50/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0067 - mae: 0.0610 - val_loss: 9.1908e-04 - val_mae: 0.0254


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
rnn_prediction = scaler_y.inverse_transform(model_rnn.predict(x_test))

#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), rnn_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (multi-layer RNN model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_2=round(mean_MAE[0],4)
MAE_val_2=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_2)
print("The validation MAE of the model is", MAE_val_2)

The MAE of the model is 0.0601
The validation MAE of the model is 0.0817


tf.keras.backend.clear_session()

# Using info New York
info_sel_state = sel_state("California")
info_train_test=train_test("California")

# Scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# minmax scaler for y, scaling between 0 and 1
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)

#Call x variables
xtrain=info_train_test[4]
xtrain.drop(["date", "JHU_cases"], axis=1, inplace=True)
xtest=info_train_test[5].copy()
xtest.drop(["date", "JHU_cases"], axis=1, inplace=True)

# time series generator uses 14 day block to predict the next day 
train_gen = TimeseriesGenerator(xtrain.to_numpy(), ytrain,
                               length=14, sampling_rate=1,  
                                batch_size = 558)

test_gen = TimeseriesGenerator(xtest.to_numpy(), ytest,
                                length=14, sampling_rate=1, 
                               batch_size = 186)

# Below we are creating x and y train and test variable from the generators
x_train, y_train = train_gen[0]
# x_val, y_val = val_gen[0]
x_test, y_test = test_gen[0]

print('x train shape is :', x_train.shape)
print('y train shape is :', y_train.shape)
print('x test shape is :' , x_test.shape)
print('y test shape is :' , y_test.shape)

x train shape is : (544, 14, 498)
y train shape is : (544, 1)
x test shape is : (172, 14, 498)
y test shape is : (172, 1)


# input dimension below
input_dim = x_train.shape[1:]
n_units = 100

#Create model
model_rnn_input = tf.keras.Input(shape=input_dim)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units, return_sequences=False)(model_rnn_input)
model_rnn_output = tf.keras.layers.Dense(units=1, activation='linear')(model_rnn_hidden)
model_rnn = tf.keras.Model(inputs=model_rnn_input, outputs=model_rnn_output, name="model_rnn")

#Print the model architecture
print(model_rnn.summary())

#Compile model
model_rnn.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_rnn.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_rnn"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 simple_rnn (SimpleRNN)      (None, 100)               59900     
                                                                 
 dense (Dense)               (None, 1)                 101       
                                                                 
=================================================================
Total params: 60,001
Trainable params: 60,001
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0399 - mae: 0.1217 - val_loss: 0.0082 - val_mae: 0.0801
Epoch 2/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0092 - mae: 0.0701 - val_loss: 0.0023 - val_mae: 0.0368
Epoch 3/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0065 - mae: 0.0560 - val_loss: 0.0091 - val_mae: 0.0864
Epoch 4/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0043 - mae: 0.0456 - val_loss: 0.0016 - val_mae: 0.0299
Epoch 5/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0049 - mae: 0.0528 - val_loss: 0.0011 - val_mae: 0.0264
Epoch 6/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0048 - mae: 0.0484 - val_loss: 0.0160 - val_mae: 0.1221
Epoch 7/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0055 - mae: 0.0512 - val_loss: 0.0246 - val_mae: 0.1550
Epoch 8/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0042 - mae: 0.0468 - val_loss: 0.0015 - val_mae: 0.0287
Epoch 9/50
489/489 [==============================] - 2s 4ms/step - loss: 0.0057 - mae: 0.0569 - val_loss: 0.0125 - val_mae: 0.1101
Epoch 10/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0064 - mae: 0.0569 - val_loss: 7.1169e-04 - val_mae: 0.0208
Epoch 11/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0045 - mae: 0.0463 - val_loss: 5.9596e-04 - val_mae: 0.0153
Epoch 12/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0051 - mae: 0.0503 - val_loss: 9.2196e-04 - val_mae: 0.0248
Epoch 13/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0062 - mae: 0.0560 - val_loss: 0.0011 - val_mae: 0.0258
Epoch 14/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0063 - mae: 0.0558 - val_loss: 0.0026 - val_mae: 0.0480
Epoch 15/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0054 - mae: 0.0536 - val_loss: 6.1158e-04 - val_mae: 0.0149
Epoch 16/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0044 - mae: 0.0477 - val_loss: 0.0060 - val_mae: 0.0749
Epoch 17/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0046 - mae: 0.0482 - val_loss: 0.0014 - val_mae: 0.0306
Epoch 18/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0038 - mae: 0.0435 - val_loss: 0.0018 - val_mae: 0.0391
Epoch 19/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0039 - mae: 0.0449 - val_loss: 9.7040e-04 - val_mae: 0.0267
Epoch 20/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0025 - mae: 0.0362 - val_loss: 0.0019 - val_mae: 0.0390
Epoch 21/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0041 - mae: 0.0423 - val_loss: 0.0032 - val_mae: 0.0535
Epoch 22/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0029 - mae: 0.0377 - val_loss: 6.6264e-04 - val_mae: 0.0177
Epoch 23/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0036 - mae: 0.0428 - val_loss: 6.9806e-04 - val_mae: 0.0194
Epoch 24/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0031 - mae: 0.0395 - val_loss: 0.0011 - val_mae: 0.0292
Epoch 25/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0032 - mae: 0.0421 - val_loss: 9.7163e-04 - val_mae: 0.0241
Epoch 26/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0021 - mae: 0.0331 - val_loss: 0.0020 - val_mae: 0.0404
Epoch 27/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0032 - mae: 0.0417 - val_loss: 0.0019 - val_mae: 0.0383
Epoch 28/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0028 - mae: 0.0388 - val_loss: 0.0018 - val_mae: 0.0368
Epoch 29/50
489/489 [==============================] - 3s 5ms/step - loss: 0.0035 - mae: 0.0429 - val_loss: 0.0021 - val_mae: 0.0436
Epoch 30/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0038 - mae: 0.0459 - val_loss: 0.0055 - val_mae: 0.0690
Epoch 31/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0024 - mae: 0.0339 - val_loss: 0.0013 - val_mae: 0.0292
Epoch 32/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0022 - mae: 0.0343 - val_loss: 0.0027 - val_mae: 0.0489
Epoch 33/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0018 - mae: 0.0297 - val_loss: 8.4147e-04 - val_mae: 0.0216
Epoch 34/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0026 - mae: 0.0364 - val_loss: 6.7294e-04 - val_mae: 0.0187
Epoch 35/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0022 - mae: 0.0359 - val_loss: 4.6460e-04 - val_mae: 0.0114
Epoch 36/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0048 - mae: 0.0496 - val_loss: 0.0026 - val_mae: 0.0480
Epoch 37/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0023 - mae: 0.0337 - val_loss: 5.8362e-04 - val_mae: 0.0157
Epoch 38/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0027 - mae: 0.0377 - val_loss: 0.0016 - val_mae: 0.0349
Epoch 39/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0019 - mae: 0.0316 - val_loss: 0.0011 - val_mae: 0.0254
Epoch 40/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0057 - mae: 0.0528 - val_loss: 0.0235 - val_mae: 0.1512
Epoch 41/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0032 - mae: 0.0426 - val_loss: 0.0018 - val_mae: 0.0377
Epoch 42/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0018 - mae: 0.0319 - val_loss: 0.0014 - val_mae: 0.0325
Epoch 43/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0018 - mae: 0.0315 - val_loss: 6.1982e-04 - val_mae: 0.0171
Epoch 44/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0025 - mae: 0.0392 - val_loss: 0.0015 - val_mae: 0.0353
Epoch 45/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0017 - mae: 0.0307 - val_loss: 0.0076 - val_mae: 0.0838
Epoch 46/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0029 - mae: 0.0410 - val_loss: 7.9019e-04 - val_mae: 0.0200
Epoch 47/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0021 - mae: 0.0340 - val_loss: 6.0138e-04 - val_mae: 0.0147
Epoch 48/50
489/489 [==============================] - 2s 4ms/step - loss: 0.0025 - mae: 0.0381 - val_loss: 6.9023e-04 - val_mae: 0.0197
Epoch 49/50
489/489 [==============================] - 2s 5ms/step - loss: 0.0021 - mae: 0.0351 - val_loss: 6.9399e-04 - val_mae: 0.0180
Epoch 50/50
489/489 [==============================] - 2s 4ms/step - loss: 0.0019 - mae: 0.0331 - val_loss: 0.0037 - val_mae: 0.0577


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of one-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of one-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
rnn_prediction = scaler_y.inverse_transform(model_rnn.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), rnn_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (one-layer RNN model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_3=round(mean_MAE[0],4)
MAE_val_3=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_3)
print("The validation MAE of the model is", MAE_val_3)

The MAE of the model is 0.0446
The validation MAE of the model is 0.043


# input dimension below
input_dim = x_train.shape[1:]
n_units = 128
n_units2 = 64

#Create model
model_rnn_input = tf.keras.Input(shape=input_dim)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units, return_sequences=True, dropout=0.2 )(model_rnn_input)
model_rnn_hidden = tf.keras.layers.SimpleRNN(units=n_units2, return_sequences=False, dropout=0.2)(model_rnn_hidden)
model_rnn_output = tf.keras.layers.Dense(units=1, activation='linear')(model_rnn_hidden)
model_rnn = tf.keras.Model(inputs=model_rnn_input, outputs=model_rnn_output, name="model_rnn")

#Print the model architecture
print(model_rnn.summary())

#Compile model
model_rnn.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_rnn.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_rnn"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 simple_rnn_1 (SimpleRNN)    (None, 14, 128)           80256     
                                                                 
 simple_rnn_2 (SimpleRNN)    (None, 64)                12352     
                                                                 
 dense_1 (Dense)             (None, 1)                 65        
                                                                 
=================================================================
Total params: 92,673
Trainable params: 92,673
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 6s 9ms/step - loss: 0.0828 - mae: 0.1981 - val_loss: 0.0043 - val_mae: 0.0634
Epoch 2/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0158 - mae: 0.0874 - val_loss: 0.0022 - val_mae: 0.0445
Epoch 3/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0105 - mae: 0.0731 - val_loss: 0.0025 - val_mae: 0.0471
Epoch 4/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0088 - mae: 0.0646 - val_loss: 9.3596e-04 - val_mae: 0.0270
Epoch 5/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0075 - mae: 0.0586 - val_loss: 0.0018 - val_mae: 0.0395
Epoch 6/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0109 - mae: 0.0737 - val_loss: 0.0031 - val_mae: 0.0533
Epoch 7/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0102 - mae: 0.0683 - val_loss: 0.0050 - val_mae: 0.0681
Epoch 8/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0104 - mae: 0.0716 - val_loss: 0.0094 - val_mae: 0.0949
Epoch 9/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0191 - mae: 0.0878 - val_loss: 0.1451 - val_mae: 0.3804
Epoch 10/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0160 - mae: 0.0839 - val_loss: 0.0012 - val_mae: 0.0325
Epoch 11/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0108 - mae: 0.0713 - val_loss: 0.0013 - val_mae: 0.0318
Epoch 12/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0081 - mae: 0.0627 - val_loss: 0.0011 - val_mae: 0.0300
Epoch 13/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0079 - mae: 0.0634 - val_loss: 0.0064 - val_mae: 0.0773
Epoch 14/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0096 - mae: 0.0687 - val_loss: 5.1149e-04 - val_mae: 0.0129
Epoch 15/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0069 - mae: 0.0568 - val_loss: 0.0016 - val_mae: 0.0364
Epoch 16/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0089 - mae: 0.0653 - val_loss: 0.0055 - val_mae: 0.0679
Epoch 17/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0068 - mae: 0.0588 - val_loss: 7.2237e-04 - val_mae: 0.0226
Epoch 18/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0083 - mae: 0.0617 - val_loss: 0.0012 - val_mae: 0.0308
Epoch 19/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0096 - mae: 0.0646 - val_loss: 0.0083 - val_mae: 0.0890
Epoch 20/50
489/489 [==============================] - 4s 9ms/step - loss: 0.0083 - mae: 0.0624 - val_loss: 7.3189e-04 - val_mae: 0.0224
Epoch 21/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0067 - mae: 0.0559 - val_loss: 6.5925e-04 - val_mae: 0.0210
Epoch 22/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0112 - mae: 0.0740 - val_loss: 0.0032 - val_mae: 0.0541
Epoch 23/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0097 - mae: 0.0696 - val_loss: 0.0073 - val_mae: 0.0832
Epoch 24/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0094 - mae: 0.0698 - val_loss: 0.0026 - val_mae: 0.0486
Epoch 25/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0084 - mae: 0.0648 - val_loss: 0.0038 - val_mae: 0.0592
Epoch 26/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0108 - mae: 0.0661 - val_loss: 0.0049 - val_mae: 0.0675
Epoch 27/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0181 - mae: 0.0907 - val_loss: 0.0088 - val_mae: 0.0916
Epoch 28/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0088 - mae: 0.0685 - val_loss: 0.0014 - val_mae: 0.0355
Epoch 29/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0096 - mae: 0.0681 - val_loss: 4.8963e-04 - val_mae: 0.0155
Epoch 30/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0067 - mae: 0.0591 - val_loss: 4.3153e-04 - val_mae: 0.0128
Epoch 31/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0072 - mae: 0.0631 - val_loss: 0.0011 - val_mae: 0.0299
Epoch 32/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0076 - mae: 0.0631 - val_loss: 0.0082 - val_mae: 0.0880
Epoch 33/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0098 - mae: 0.0680 - val_loss: 9.0314e-04 - val_mae: 0.0266
Epoch 34/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0107 - mae: 0.0728 - val_loss: 6.7268e-04 - val_mae: 0.0207
Epoch 35/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0102 - mae: 0.0713 - val_loss: 4.1960e-04 - val_mae: 0.0102
Epoch 36/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0072 - mae: 0.0611 - val_loss: 0.0024 - val_mae: 0.0464
Epoch 37/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0118 - mae: 0.0756 - val_loss: 0.0049 - val_mae: 0.0673
Epoch 38/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0121 - mae: 0.0803 - val_loss: 0.0031 - val_mae: 0.0533
Epoch 39/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0121 - mae: 0.0815 - val_loss: 0.0250 - val_mae: 0.1569
Epoch 40/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0105 - mae: 0.0741 - val_loss: 0.0017 - val_mae: 0.0383
Epoch 41/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0089 - mae: 0.0676 - val_loss: 6.2606e-04 - val_mae: 0.0174
Epoch 42/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0117 - mae: 0.0762 - val_loss: 4.3084e-04 - val_mae: 0.0104
Epoch 43/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0129 - mae: 0.0760 - val_loss: 7.5056e-04 - val_mae: 0.0233
Epoch 44/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0067 - mae: 0.0595 - val_loss: 0.0069 - val_mae: 0.0805
Epoch 45/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0083 - mae: 0.0657 - val_loss: 7.2243e-04 - val_mae: 0.0212
Epoch 46/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0093 - mae: 0.0657 - val_loss: 0.0018 - val_mae: 0.0372
Epoch 47/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0062 - mae: 0.0554 - val_loss: 0.0016 - val_mae: 0.0371
Epoch 48/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0129 - mae: 0.0760 - val_loss: 0.0062 - val_mae: 0.0763
Epoch 49/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0109 - mae: 0.0720 - val_loss: 0.0020 - val_mae: 0.0421
Epoch 50/50
489/489 [==============================] - 4s 8ms/step - loss: 0.0066 - mae: 0.0600 - val_loss: 0.0059 - val_mae: 0.0742


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer RNN model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
rnn_prediction = scaler_y.inverse_transform(model_rnn.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), rnn_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (multi-layer RNN model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_4=round(mean_MAE[0],4)
MAE_val_4=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_4)
print("The validation MAE of the model is", MAE_val_4)

The MAE of the model is 0.0715
The validation MAE of the model is 0.0544


tf.keras.backend.clear_session()

# Using info New York
info_sel_state = sel_state("New York")
info_train_test=train_test("New York")

# Scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# minmax scaler for y, scaling between 0 and 1
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)

#Call x variables
xtrain=info_train_test[4]
xtrain.drop(["date", "JHU_cases"], axis=1, inplace=True)
xtest=info_train_test[5].copy()
xtest.drop(["date", "JHU_cases"], axis=1, inplace=True)

# time series generator uses 14 day block to predict the next day 
train_gen = TimeseriesGenerator(xtrain.to_numpy(), ytrain,
                               length=14, sampling_rate=1,  
                                batch_size = 558)

test_gen = TimeseriesGenerator(xtest.to_numpy(), ytest,
                                length=14, sampling_rate=1, 
                               batch_size = 186)

# Below we are creating x and y train and test variable from the generators
x_train, y_train = train_gen[0]
# x_val, y_val = val_gen[0]
x_test, y_test = test_gen[0]

print('x train shape is :', x_train.shape)
print('y train shape is :', y_train.shape)
print('x test shape is :' , x_test.shape)
print('y test shape is :' , y_test.shape)

x train shape is : (544, 14, 498)
y train shape is : (544, 1)
x test shape is : (172, 14, 498)
y test shape is : (172, 1)


# input dimension below
input_dim = x_train.shape[1:]
n_units = 100

#Create model
model_lstm_input = tf.keras.Input(shape=input_dim)
model_hidden=tf.keras.layers.LSTM(units = n_units)(model_lstm_input)
model_lstm_output=tf.keras.layers.Dense(units = 1, activation="linear")(model_hidden)
model_lstm = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm")


#Print the model architecture
print(model_lstm.summary())

#Compile model
model_lstm.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_lstm.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_lstm"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 100)               239600    
                                                                 
 dense (Dense)               (None, 1)                 101       
                                                                 
=================================================================
Total params: 239,701
Trainable params: 239,701
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 9s 14ms/step - loss: 0.0278 - mae: 0.0851 - val_loss: 0.0233 - val_mae: 0.1460
Epoch 2/50
489/489 [==============================] - 6s 13ms/step - loss: 0.0044 - mae: 0.0436 - val_loss: 0.0032 - val_mae: 0.0474
Epoch 3/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0034 - mae: 0.0374 - val_loss: 0.0062 - val_mae: 0.0737
Epoch 4/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0033 - mae: 0.0383 - val_loss: 0.0031 - val_mae: 0.0500
Epoch 5/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0037 - mae: 0.0403 - val_loss: 0.0065 - val_mae: 0.0765
Epoch 6/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0025 - mae: 0.0299 - val_loss: 0.0156 - val_mae: 0.1212
Epoch 7/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0024 - mae: 0.0319 - val_loss: 0.0015 - val_mae: 0.0301
Epoch 8/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0022 - mae: 0.0307 - val_loss: 0.0042 - val_mae: 0.0584
Epoch 9/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0021 - mae: 0.0323 - val_loss: 0.0032 - val_mae: 0.0497
Epoch 10/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0020 - mae: 0.0303 - val_loss: 8.7140e-04 - val_mae: 0.0221
Epoch 11/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0020 - mae: 0.0294 - val_loss: 0.0041 - val_mae: 0.0585
Epoch 12/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0018 - mae: 0.0269 - val_loss: 0.0058 - val_mae: 0.0718
Epoch 13/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0017 - mae: 0.0282 - val_loss: 0.0015 - val_mae: 0.0307
Epoch 14/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0014 - mae: 0.0262 - val_loss: 0.0026 - val_mae: 0.0443
Epoch 15/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0014 - mae: 0.0249 - val_loss: 0.0031 - val_mae: 0.0480
Epoch 16/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0012 - mae: 0.0252 - val_loss: 0.0042 - val_mae: 0.0593
Epoch 17/50
489/489 [==============================] - 6s 12ms/step - loss: 9.7890e-04 - mae: 0.0233 - val_loss: 0.0029 - val_mae: 0.0458
Epoch 18/50
489/489 [==============================] - 6s 12ms/step - loss: 8.3854e-04 - mae: 0.0211 - val_loss: 0.0032 - val_mae: 0.0519
Epoch 19/50
489/489 [==============================] - 6s 12ms/step - loss: 9.5015e-04 - mae: 0.0224 - val_loss: 0.0014 - val_mae: 0.0297
Epoch 20/50
489/489 [==============================] - 6s 12ms/step - loss: 8.4273e-04 - mae: 0.0210 - val_loss: 0.0026 - val_mae: 0.0440
Epoch 21/50
489/489 [==============================] - 6s 12ms/step - loss: 4.6586e-04 - mae: 0.0165 - val_loss: 0.0013 - val_mae: 0.0285
Epoch 22/50
489/489 [==============================] - 6s 12ms/step - loss: 9.0573e-04 - mae: 0.0210 - val_loss: 0.0040 - val_mae: 0.0592
Epoch 23/50
489/489 [==============================] - 6s 12ms/step - loss: 8.1101e-04 - mae: 0.0209 - val_loss: 0.0034 - val_mae: 0.0540
Epoch 24/50
489/489 [==============================] - 6s 13ms/step - loss: 5.1372e-04 - mae: 0.0164 - val_loss: 0.0038 - val_mae: 0.0582
Epoch 25/50
489/489 [==============================] - 6s 13ms/step - loss: 6.1445e-04 - mae: 0.0182 - val_loss: 0.0035 - val_mae: 0.0547
Epoch 26/50
489/489 [==============================] - 6s 12ms/step - loss: 4.7087e-04 - mae: 0.0163 - val_loss: 0.0021 - val_mae: 0.0407
Epoch 27/50
489/489 [==============================] - 6s 12ms/step - loss: 7.0810e-04 - mae: 0.0189 - val_loss: 0.0015 - val_mae: 0.0321
Epoch 28/50
489/489 [==============================] - 6s 12ms/step - loss: 3.4888e-04 - mae: 0.0141 - val_loss: 0.0018 - val_mae: 0.0372
Epoch 29/50
489/489 [==============================] - 6s 13ms/step - loss: 3.6345e-04 - mae: 0.0146 - val_loss: 0.0032 - val_mae: 0.0524
Epoch 30/50
489/489 [==============================] - 6s 13ms/step - loss: 3.9726e-04 - mae: 0.0147 - val_loss: 0.0038 - val_mae: 0.0581
Epoch 31/50
489/489 [==============================] - 6s 13ms/step - loss: 5.6201e-04 - mae: 0.0168 - val_loss: 0.0041 - val_mae: 0.0605
Epoch 32/50
489/489 [==============================] - 6s 13ms/step - loss: 4.3212e-04 - mae: 0.0152 - val_loss: 5.6658e-04 - val_mae: 0.0180
Epoch 33/50
489/489 [==============================] - 6s 13ms/step - loss: 4.8134e-04 - mae: 0.0158 - val_loss: 0.0021 - val_mae: 0.0427
Epoch 34/50
489/489 [==============================] - 6s 13ms/step - loss: 3.9820e-04 - mae: 0.0145 - val_loss: 0.0019 - val_mae: 0.0389
Epoch 35/50
489/489 [==============================] - 6s 13ms/step - loss: 2.9571e-04 - mae: 0.0128 - val_loss: 0.0013 - val_mae: 0.0314
Epoch 36/50
489/489 [==============================] - 6s 12ms/step - loss: 3.5831e-04 - mae: 0.0140 - val_loss: 0.0012 - val_mae: 0.0296
Epoch 37/50
489/489 [==============================] - 6s 12ms/step - loss: 2.4895e-04 - mae: 0.0119 - val_loss: 0.0011 - val_mae: 0.0282
Epoch 38/50
489/489 [==============================] - 6s 13ms/step - loss: 5.5265e-04 - mae: 0.0162 - val_loss: 9.3092e-04 - val_mae: 0.0261
Epoch 39/50
489/489 [==============================] - 6s 13ms/step - loss: 3.0301e-04 - mae: 0.0128 - val_loss: 0.0037 - val_mae: 0.0577
Epoch 40/50
489/489 [==============================] - 6s 12ms/step - loss: 2.4097e-04 - mae: 0.0115 - val_loss: 0.0031 - val_mae: 0.0523
Epoch 41/50
489/489 [==============================] - 6s 13ms/step - loss: 3.1027e-04 - mae: 0.0121 - val_loss: 0.0029 - val_mae: 0.0505
Epoch 42/50
489/489 [==============================] - 6s 13ms/step - loss: 4.0832e-04 - mae: 0.0138 - val_loss: 0.0013 - val_mae: 0.0319
Epoch 43/50
489/489 [==============================] - 6s 12ms/step - loss: 1.7997e-04 - mae: 0.0099 - val_loss: 0.0029 - val_mae: 0.0510
Epoch 44/50
489/489 [==============================] - 6s 12ms/step - loss: 1.7004e-04 - mae: 0.0098 - val_loss: 0.0022 - val_mae: 0.0435
Epoch 45/50
489/489 [==============================] - 6s 12ms/step - loss: 2.7693e-04 - mae: 0.0122 - val_loss: 0.0020 - val_mae: 0.0403
Epoch 46/50
489/489 [==============================] - 6s 12ms/step - loss: 3.3876e-04 - mae: 0.0132 - val_loss: 0.0042 - val_mae: 0.0623
Epoch 47/50
489/489 [==============================] - 6s 12ms/step - loss: 2.1596e-04 - mae: 0.0107 - val_loss: 0.0019 - val_mae: 0.0404
Epoch 48/50
489/489 [==============================] - 6s 13ms/step - loss: 3.8830e-04 - mae: 0.0131 - val_loss: 0.0024 - val_mae: 0.0453
Epoch 49/50
489/489 [==============================] - 6s 13ms/step - loss: 2.6645e-04 - mae: 0.0115 - val_loss: 0.0055 - val_mae: 0.0711
Epoch 50/50
489/489 [==============================] - 6s 12ms/step - loss: 4.7244e-04 - mae: 0.0143 - val_loss: 0.0016 - val_mae: 0.0362


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of one-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of one-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
lstm_prediction = scaler_y.inverse_transform(model_lstm.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), lstm_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (one-layer LSTM model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_5=round(mean_MAE[0],4)
MAE_val_5=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_5)
print("The validation MAE of the model is", MAE_val_5)

The MAE of the model is 0.0216
The validation MAE of the model is 0.0498


### Multilayer LSTM model
#Design following: Iberoamerican Journal of Medicine
tf.keras.backend.clear_session()

# input dimension below
input_dim = x_train.shape[1:]
n_units= 50

#Create model
model_lstm_input = tf.keras.Input(shape=input_dim)
model_hidden1=tf.keras.layers.LSTM(units = n_units, return_sequences=True)(model_lstm_input)
model_dropout1=tf.keras.layers.Dropout(rate=0.2)(model_hidden1)
model_hidden2=tf.keras.layers.LSTM(units = n_units, return_sequences=True)(model_dropout1)
model_dropout2=tf.keras.layers.Dropout(rate=0.2)(model_hidden2)
model_hidden3=tf.keras.layers.LSTM(units = n_units, return_sequences=False)(model_dropout2)
model_dropout3=tf.keras.layers.Dropout(rate=0.2)(model_hidden3)
model_lstm_output=tf.keras.layers.Dense(units = 1, activation="linear")(model_dropout3)
model_lstm = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm")

#Print the model architecture
print(model_lstm.summary())

#Compile model
model_lstm.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_lstm.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_lstm"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 50)            109800    
                                                                 
 dropout (Dropout)           (None, 14, 50)            0         
                                                                 
 lstm_1 (LSTM)               (None, 14, 50)            20200     
                                                                 
 dropout_1 (Dropout)         (None, 14, 50)            0         
                                                                 
 lstm_2 (LSTM)               (None, 50)                20200     
                                                                 
 dropout_2 (Dropout)         (None, 50)                0         
                                                                 
 dense (Dense)               (None, 1)                 51        
                                                                 
=================================================================
Total params: 150,251
Trainable params: 150,251
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 16s 22ms/step - loss: 0.0129 - mae: 0.0810 - val_loss: 0.0074 - val_mae: 0.0856
Epoch 2/50
489/489 [==============================] - 9s 17ms/step - loss: 0.0074 - mae: 0.0566 - val_loss: 0.0071 - val_mae: 0.0838
Epoch 3/50
489/489 [==============================] - 9s 17ms/step - loss: 0.0056 - mae: 0.0477 - val_loss: 0.0208 - val_mae: 0.1440
Epoch 4/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0052 - mae: 0.0450 - val_loss: 0.0050 - val_mae: 0.0705
Epoch 5/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0049 - mae: 0.0446 - val_loss: 0.0061 - val_mae: 0.0779
Epoch 6/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0044 - mae: 0.0407 - val_loss: 0.0078 - val_mae: 0.0879
Epoch 7/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0046 - mae: 0.0426 - val_loss: 0.0117 - val_mae: 0.1074
Epoch 8/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0039 - mae: 0.0377 - val_loss: 0.0025 - val_mae: 0.0492
Epoch 9/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0045 - mae: 0.0426 - val_loss: 0.0052 - val_mae: 0.0716
Epoch 10/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0039 - mae: 0.0373 - val_loss: 0.0023 - val_mae: 0.0472
Epoch 11/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0042 - mae: 0.0394 - val_loss: 7.9464e-05 - val_mae: 0.0072
Epoch 12/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0042 - mae: 0.0401 - val_loss: 0.0011 - val_mae: 0.0320
Epoch 13/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0034 - mae: 0.0368 - val_loss: 0.0074 - val_mae: 0.0856
Epoch 14/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0039 - mae: 0.0378 - val_loss: 0.0019 - val_mae: 0.0422
Epoch 15/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0031 - mae: 0.0324 - val_loss: 2.5212e-04 - val_mae: 0.0141
Epoch 16/50
489/489 [==============================] - 9s 17ms/step - loss: 0.0037 - mae: 0.0387 - val_loss: 0.0024 - val_mae: 0.0483
Epoch 17/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0038 - mae: 0.0387 - val_loss: 0.0046 - val_mae: 0.0670
Epoch 18/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0029 - mae: 0.0328 - val_loss: 0.0063 - val_mae: 0.0785
Epoch 19/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0032 - mae: 0.0349 - val_loss: 0.0068 - val_mae: 0.0815
Epoch 20/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0033 - mae: 0.0362 - val_loss: 0.0088 - val_mae: 0.0921
Epoch 21/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0030 - mae: 0.0332 - val_loss: 0.0034 - val_mae: 0.0566
Epoch 22/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0029 - mae: 0.0319 - val_loss: 0.0018 - val_mae: 0.0407
Epoch 23/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0030 - mae: 0.0330 - val_loss: 0.0211 - val_mae: 0.1433
Epoch 24/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0029 - mae: 0.0334 - val_loss: 5.1619e-04 - val_mae: 0.0207
Epoch 25/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0033 - mae: 0.0352 - val_loss: 2.2611e-04 - val_mae: 0.0133
Epoch 26/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0029 - mae: 0.0334 - val_loss: 7.4860e-04 - val_mae: 0.0255
Epoch 27/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0028 - mae: 0.0325 - val_loss: 0.0040 - val_mae: 0.0619
Epoch 28/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0024 - mae: 0.0288 - val_loss: 7.3414e-04 - val_mae: 0.0254
Epoch 29/50
489/489 [==============================] - 9s 17ms/step - loss: 0.0031 - mae: 0.0350 - val_loss: 4.8799e-04 - val_mae: 0.0203
Epoch 30/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0028 - mae: 0.0321 - val_loss: 5.7452e-04 - val_mae: 0.0221
Epoch 31/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0314 - val_loss: 0.0024 - val_mae: 0.0473
Epoch 32/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0306 - val_loss: 0.0019 - val_mae: 0.0418
Epoch 33/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0023 - mae: 0.0283 - val_loss: 0.0021 - val_mae: 0.0437
Epoch 34/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0023 - mae: 0.0301 - val_loss: 0.0025 - val_mae: 0.0491
Epoch 35/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0321 - val_loss: 0.0015 - val_mae: 0.0377
Epoch 36/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0321 - val_loss: 1.5012e-04 - val_mae: 0.0081
Epoch 37/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0324 - val_loss: 0.0012 - val_mae: 0.0330
Epoch 38/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0022 - mae: 0.0281 - val_loss: 6.7038e-04 - val_mae: 0.0247
Epoch 39/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0030 - mae: 0.0328 - val_loss: 6.1319e-04 - val_mae: 0.0233
Epoch 40/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0025 - mae: 0.0302 - val_loss: 0.0021 - val_mae: 0.0446
Epoch 41/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0024 - mae: 0.0295 - val_loss: 0.0015 - val_mae: 0.0384
Epoch 42/50
489/489 [==============================] - 9s 17ms/step - loss: 0.0025 - mae: 0.0284 - val_loss: 7.0564e-04 - val_mae: 0.0249
Epoch 43/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0024 - mae: 0.0284 - val_loss: 5.2689e-04 - val_mae: 0.0213
Epoch 44/50
489/489 [==============================] - 9s 18ms/step - loss: 0.0026 - mae: 0.0315 - val_loss: 0.0041 - val_mae: 0.0626
Epoch 45/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0024 - mae: 0.0301 - val_loss: 7.1488e-04 - val_mae: 0.0252
Epoch 46/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0023 - mae: 0.0290 - val_loss: 4.5437e-04 - val_mae: 0.0195
Epoch 47/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0020 - mae: 0.0258 - val_loss: 0.0010 - val_mae: 0.0301
Epoch 48/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0015 - mae: 0.0258 - val_loss: 8.8815e-04 - val_mae: 0.0282
Epoch 49/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0025 - mae: 0.0307 - val_loss: 0.0016 - val_mae: 0.0381
Epoch 50/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0019 - mae: 0.0266 - val_loss: 5.8761e-04 - val_mae: 0.0222


# Plotting the loss and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
lstm_prediction = scaler_y.inverse_transform(model_lstm.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), lstm_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_6=round(mean_MAE[0],4)
MAE_val_6=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_6)
print("The validation MAE of the model is", MAE_val_6)

The MAE of the model is 0.0353
The validation MAE of the model is 0.0493


# Function to create datasets for LSTM as time steps
def format_series(n_future,n_past,df):
  # Creates series of multi-step on the dataset based on n_past periods
  # and adds the corresponding predicted values based on n_future periods
    trainX=[]
    trainY=[]
    for i in range(n_past,len(df)-n_future+1):
        trainX.append(df.iloc[i-n_past:i,1:df.shape[1]])#All features except y
        trainY.append(df.iloc[i+n_future-1:i+n_future,0])#y

    return np.array(trainX),np.array(trainY)


#Using info New York
info_sel_state = sel_state("New York")
info_train_test=train_test("New York")

#x and y train and test variables
x_train=info_train_test[4]
x_train.drop(["date"], axis=1, inplace=True)
y_train=info_train_test[1]
x_test=info_train_test[5]
x_test.drop(["date"], axis=1, inplace=True)
y_test=info_train_test[3]

# Datasets shape
print("\nThe shape of xtrain is", x_train.shape)
print("The shape of ytrain is", y_train.shape)

print("\nThe shape of xtest is", x_test.shape)
print("The shape of ytest is", y_test.shape)

The shape of xtrain is (558, 499)
The shape of ytrain is (558,)

The shape of xtest is (186, 499)
The shape of ytest is (186,)


# Generating Time Step Series
xtrain,ytrain = format_series(1,14,x_train) 
xtest,ytest = format_series(1,14,x_test) 

# Checking the test and train data shapes
print("\nThe shape of xtrain is", xtrain.shape)
print("The shape of ytrain is", ytrain.shape)

print("\nThe shape of xtest is", xtest.shape)
print("The shape of ytest is", ytest.shape)

The shape of xtrain is (544, 14, 498)
The shape of ytrain is (544, 1)

The shape of xtest is (172, 14, 498)
The shape of ytest is (172, 1)


tf.keras.backend.clear_session()

n_units=64

# Model architecture with 2 LSTM layers
model_lstm_input = tf.keras.Input(shape=(xtrain.shape[1], xtrain.shape[2]))
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_lstm_input)
model_hidden=tf.keras.layers.LSTM(units = n_units//2, activation='relu', return_sequences=False)(model_hidden)
model_hidden=tf.keras.layers.Dropout(0.2)(model_hidden)
model_lstm_output=tf.keras.layers.Dense(units = 1)(model_hidden)

model_lstm_shift = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm_time_steps")

#Print the model architecture
print(model_lstm_shift.summary())

#Compile model
model_lstm_shift.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Fit model
model_lstm_shift_results=model_lstm_shift.fit(xtrain, ytrain, 
                                               validation_split=0.2, 
                                               epochs = 20,
                                               batch_size = 32)

#Predict with model
model_lstm_shift_predict=model_lstm_shift.predict(xtest)

Model: "model_lstm_time_steps"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 64)            144128    
                                                                 
 lstm_1 (LSTM)               (None, 32)                12416     
                                                                 
 dropout (Dropout)           (None, 32)                0         
                                                                 
 dense (Dense)               (None, 1)                 33        
                                                                 
=================================================================
Total params: 156,577
Trainable params: 156,577
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/20
14/14 [==============================] - 3s 65ms/step - loss: 38032584.0000 - mae: 4083.8735 - val_loss: 11794229.0000 - val_mae: 2279.8066
Epoch 2/20
14/14 [==============================] - 0s 34ms/step - loss: 28557522.0000 - mae: 3573.7761 - val_loss: 12371511.0000 - val_mae: 3248.2581
Epoch 3/20
14/14 [==============================] - 0s 32ms/step - loss: 20461796.0000 - mae: 3581.7529 - val_loss: 16567672.0000 - val_mae: 3728.1582
Epoch 4/20
14/14 [==============================] - 0s 34ms/step - loss: 21745238.0000 - mae: 3834.3528 - val_loss: 24279820.0000 - val_mae: 4448.5068
Epoch 5/20
14/14 [==============================] - 0s 32ms/step - loss: 19951934.0000 - mae: 3684.2104 - val_loss: 28846816.0000 - val_mae: 4912.1665
Epoch 6/20
14/14 [==============================] - 0s 32ms/step - loss: 19773254.0000 - mae: 3657.8945 - val_loss: 13264450.0000 - val_mae: 3331.1821
Epoch 7/20
14/14 [==============================] - 0s 32ms/step - loss: 19561540.0000 - mae: 3442.4448 - val_loss: 22752744.0000 - val_mae: 4339.6514
Epoch 8/20
14/14 [==============================] - 0s 31ms/step - loss: 19618878.0000 - mae: 3611.3650 - val_loss: 16545329.0000 - val_mae: 3742.4451
Epoch 9/20
14/14 [==============================] - 0s 35ms/step - loss: 18294454.0000 - mae: 3458.4534 - val_loss: 14886128.0000 - val_mae: 3564.4922
Epoch 10/20
14/14 [==============================] - 0s 33ms/step - loss: 17201998.0000 - mae: 3161.6948 - val_loss: 9109334.0000 - val_mae: 2713.2698
Epoch 11/20
14/14 [==============================] - 0s 33ms/step - loss: 17597950.0000 - mae: 3079.6040 - val_loss: 26253044.0000 - val_mae: 4687.2319
Epoch 12/20
14/14 [==============================] - 0s 32ms/step - loss: 15511091.0000 - mae: 3206.5676 - val_loss: 16536207.0000 - val_mae: 3757.0591
Epoch 13/20
14/14 [==============================] - 0s 34ms/step - loss: 12338746.0000 - mae: 2518.1926 - val_loss: 5467000.0000 - val_mae: 2008.2953
Epoch 14/20
14/14 [==============================] - 0s 32ms/step - loss: 10940053.0000 - mae: 2020.4807 - val_loss: 7396951.5000 - val_mae: 2096.9988
Epoch 15/20
14/14 [==============================] - 0s 30ms/step - loss: 13998703.0000 - mae: 2606.4617 - val_loss: 7911093.5000 - val_mae: 2098.1912
Epoch 16/20
14/14 [==============================] - 0s 32ms/step - loss: 21096394.0000 - mae: 3092.3257 - val_loss: 11157249.0000 - val_mae: 3093.6624
Epoch 17/20
14/14 [==============================] - 0s 32ms/step - loss: 16157805.0000 - mae: 3178.9932 - val_loss: 18118036.0000 - val_mae: 3902.0889
Epoch 18/20
14/14 [==============================] - 0s 32ms/step - loss: 12782654.0000 - mae: 2657.1682 - val_loss: 10892875.0000 - val_mae: 3050.3984
Epoch 19/20
14/14 [==============================] - 0s 31ms/step - loss: 17097910.0000 - mae: 3306.7432 - val_loss: 8439723.0000 - val_mae: 2542.4348
Epoch 20/20
14/14 [==============================] - 0s 33ms/step - loss: 17280198.0000 - mae: 2764.5601 - val_loss: 8830552.0000 - val_mae: 2760.7988


# Plot training results
epochs=20
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of two-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of two-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


#Plot actual vs. test and validation prediction
shift = 14
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.title(f"Predicted versus actual JHU_cases (two-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


# Forecasting 14 days beyond our dataset
n_future=14
forecast = model_lstm_shift.predict(xtrain[-n_future:])


#Plot actual vs. test and validation prediction
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.plot(range(len(y_train)+len(y_test),len(y_train)+len(y_test)+len(forecast)),forecast,color='r',alpha=.8,lw=2,label="Forecast")
plt.title(f"Predicted versus actual JHU_cases (two-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_7=round(mean_MAE[0],4)
MAE_val_7=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_7)
print("The validation MAE of the model is", MAE_val_7)

The MAE of the model is 0.0353
The validation MAE of the model is 0.0493


#Using info New York
info_sel_state = sel_state("New York")
info_train_test=train_test("New York")

#x and y train and test variables
x_train=info_train_test[4]
x_train.drop(["date"], axis=1, inplace=True)
y_train=info_train_test[1]
x_test=info_train_test[5]
x_test.drop(["date"], axis=1, inplace=True)
y_test=info_train_test[3]

# Datasets shape
print("\nThe shape of xtrain is", x_train.shape)
print("The shape of ytrain is", y_train.shape)

print("\nThe shape of xtest is", x_test.shape)
print("The shape of ytest is", y_test.shape)

The shape of xtrain is (558, 499)
The shape of ytrain is (558,)

The shape of xtest is (186, 499)
The shape of ytest is (186,)


# Generating Time Step Series
xtrain,ytrain = format_series(1,14,x_train) 
xtest,ytest = format_series(1,14,x_test) 

# Checking the test and train data shapes
print("\nThe shape of xtrain is", xtrain.shape)
print("The shape of ytrain is", ytrain.shape)

print("\nThe shape of xtest is", xtest.shape)
print("The shape of ytest is", ytest.shape)

The shape of xtrain is (544, 14, 498)
The shape of ytrain is (544, 1)

The shape of xtest is (172, 14, 498)
The shape of ytest is (172, 1)


tf.keras.backend.clear_session()

n_units=64

# Model architecture with 4 LSTM layers
model_lstm_input = tf.keras.Input(shape=(xtrain.shape[1], xtrain.shape[2]))
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_lstm_input)
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_hidden)
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_hidden)
model_hidden=tf.keras.layers.LSTM(units = n_units//2, activation='relu', return_sequences=False)(model_hidden)
model_hidden=tf.keras.layers.Dropout(0.2)(model_hidden)
model_lstm_output=tf.keras.layers.Dense(units = 1)(model_hidden)

model_lstm_shift = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm_time_steps")

#Print the model architecture
print(model_lstm_shift.summary())

#Compile model
model_lstm_shift.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Fit model
model_lstm_shift_results=model_lstm_shift.fit(xtrain, ytrain, 
                                               validation_split=0.2, 
                                               epochs = 50,
                                               batch_size = 32)

#Predict with model
model_lstm_shift_predict=model_lstm_shift.predict(xtest)

Model: "model_lstm_time_steps"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 64)            144128    
                                                                 
 lstm_1 (LSTM)               (None, 14, 64)            33024     
                                                                 
 lstm_2 (LSTM)               (None, 14, 64)            33024     
                                                                 
 lstm_3 (LSTM)               (None, 32)                12416     
                                                                 
 dropout (Dropout)           (None, 32)                0         
                                                                 
 dense (Dense)               (None, 1)                 33        
                                                                 
=================================================================
Total params: 222,625
Trainable params: 222,625
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
14/14 [==============================] - 7s 97ms/step - loss: 38260176.0000 - mae: 4093.9268 - val_loss: 10088403.0000 - val_mae: 2235.2007
Epoch 2/50
14/14 [==============================] - 1s 47ms/step - loss: 26615236.0000 - mae: 3610.0017 - val_loss: 14852836.0000 - val_mae: 3550.7310
Epoch 3/50
14/14 [==============================] - 1s 51ms/step - loss: 20209262.0000 - mae: 3479.1252 - val_loss: 23370210.0000 - val_mae: 4385.4644
Epoch 4/50
14/14 [==============================] - 1s 47ms/step - loss: 18983208.0000 - mae: 3433.6384 - val_loss: 13503226.0000 - val_mae: 3404.5608
Epoch 5/50
14/14 [==============================] - 1s 48ms/step - loss: 17582030.0000 - mae: 3218.2012 - val_loss: 14992870.0000 - val_mae: 3569.0183
Epoch 6/50
14/14 [==============================] - 1s 48ms/step - loss: 17928366.0000 - mae: 3198.0312 - val_loss: 13957928.0000 - val_mae: 3469.9272
Epoch 7/50
14/14 [==============================] - 1s 49ms/step - loss: 10245148.0000 - mae: 2295.8137 - val_loss: 12289368.0000 - val_mae: 3287.1458
Epoch 8/50
14/14 [==============================] - 1s 48ms/step - loss: 6821453.0000 - mae: 1470.8976 - val_loss: 4150058.0000 - val_mae: 1617.7141
Epoch 9/50
14/14 [==============================] - 1s 46ms/step - loss: 6147211.0000 - mae: 1367.9806 - val_loss: 4609080.5000 - val_mae: 1901.9875
Epoch 10/50
14/14 [==============================] - 1s 48ms/step - loss: 5579849.0000 - mae: 1318.6981 - val_loss: 5509693.0000 - val_mae: 2163.4243
Epoch 11/50
14/14 [==============================] - 1s 50ms/step - loss: 5328524.5000 - mae: 1390.7135 - val_loss: 4634629.5000 - val_mae: 1943.6030
Epoch 12/50
14/14 [==============================] - 1s 49ms/step - loss: 4740514.5000 - mae: 1216.5298 - val_loss: 5096843.5000 - val_mae: 2029.6202
Epoch 13/50
14/14 [==============================] - 1s 50ms/step - loss: 7136721.0000 - mae: 1604.1820 - val_loss: 9786974.0000 - val_mae: 2874.0867
Epoch 14/50
14/14 [==============================] - 1s 47ms/step - loss: 12919749.0000 - mae: 2315.4888 - val_loss: 30196772.0000 - val_mae: 4763.6455
Epoch 15/50
14/14 [==============================] - 1s 46ms/step - loss: 14834348.0000 - mae: 2595.7129 - val_loss: 9373637.0000 - val_mae: 2838.5596
Epoch 16/50
14/14 [==============================] - 1s 49ms/step - loss: 19098640.0000 - mae: 2991.3137 - val_loss: 10141463.0000 - val_mae: 2948.0474
Epoch 17/50
14/14 [==============================] - 1s 48ms/step - loss: 17485984.0000 - mae: 3218.7532 - val_loss: 17431428.0000 - val_mae: 3843.1360
Epoch 18/50
14/14 [==============================] - 1s 46ms/step - loss: 14997593.0000 - mae: 3106.2969 - val_loss: 16050512.0000 - val_mae: 3698.2439
Epoch 19/50
14/14 [==============================] - 1s 48ms/step - loss: 15226133.0000 - mae: 2939.2229 - val_loss: 17732448.0000 - val_mae: 3887.6272
Epoch 20/50
14/14 [==============================] - 1s 48ms/step - loss: 13328094.0000 - mae: 2732.2529 - val_loss: 10496483.0000 - val_mae: 2178.5332
Epoch 21/50
14/14 [==============================] - 1s 47ms/step - loss: 11334106.0000 - mae: 2347.7437 - val_loss: 19704482.0000 - val_mae: 3587.8381
Epoch 22/50
14/14 [==============================] - 1s 49ms/step - loss: 8779215.0000 - mae: 1955.3296 - val_loss: 4453819.0000 - val_mae: 1782.2784
Epoch 23/50
14/14 [==============================] - 1s 48ms/step - loss: 9022253.0000 - mae: 1811.5559 - val_loss: 7766824.0000 - val_mae: 2069.0439
Epoch 24/50
14/14 [==============================] - 1s 47ms/step - loss: 18458076.0000 - mae: 2701.0137 - val_loss: 7999447.0000 - val_mae: 2228.7947
Epoch 25/50
14/14 [==============================] - 1s 48ms/step - loss: 13085408.0000 - mae: 2429.9890 - val_loss: 7980085.5000 - val_mae: 2627.2168
Epoch 26/50
14/14 [==============================] - 1s 47ms/step - loss: 10908625.0000 - mae: 2378.9932 - val_loss: 8816103.0000 - val_mae: 2794.2283
Epoch 27/50
14/14 [==============================] - 1s 47ms/step - loss: 9974600.0000 - mae: 2252.8386 - val_loss: 5404559.5000 - val_mae: 2084.3638
Epoch 28/50
14/14 [==============================] - 1s 50ms/step - loss: 6606811.5000 - mae: 1645.1766 - val_loss: 4724673.0000 - val_mae: 1978.2881
Epoch 29/50
14/14 [==============================] - 1s 48ms/step - loss: 7042011.0000 - mae: 1496.2184 - val_loss: 5747079.0000 - val_mae: 2234.5862
Epoch 30/50
14/14 [==============================] - 1s 50ms/step - loss: 6577869.0000 - mae: 1479.6544 - val_loss: 8019064.0000 - val_mae: 2548.5120
Epoch 31/50
14/14 [==============================] - 1s 48ms/step - loss: 6488319.0000 - mae: 1525.8337 - val_loss: 7231746.5000 - val_mae: 2348.6799
Epoch 32/50
14/14 [==============================] - 1s 49ms/step - loss: 6202960.5000 - mae: 1407.8066 - val_loss: 1694594.7500 - val_mae: 930.1036
Epoch 33/50
14/14 [==============================] - 1s 47ms/step - loss: 5828259.5000 - mae: 1390.6769 - val_loss: 8115519.0000 - val_mae: 2584.6951
Epoch 34/50
14/14 [==============================] - 1s 54ms/step - loss: 6215967.0000 - mae: 1363.8821 - val_loss: 7565009.0000 - val_mae: 2365.7307
Epoch 35/50
14/14 [==============================] - 1s 51ms/step - loss: 5783503.5000 - mae: 1347.8904 - val_loss: 10721657.0000 - val_mae: 3080.4885
Epoch 36/50
14/14 [==============================] - 1s 47ms/step - loss: 5836080.5000 - mae: 1466.5253 - val_loss: 2569827.0000 - val_mae: 1125.6656
Epoch 37/50
14/14 [==============================] - 1s 47ms/step - loss: 6593006.0000 - mae: 1447.4344 - val_loss: 3288794.5000 - val_mae: 1412.8909
Epoch 38/50
14/14 [==============================] - 1s 48ms/step - loss: 6870197.5000 - mae: 1331.1248 - val_loss: 7056457.5000 - val_mae: 2281.2322
Epoch 39/50
14/14 [==============================] - 1s 46ms/step - loss: 5138612.0000 - mae: 1335.5691 - val_loss: 4156475.5000 - val_mae: 1784.1935
Epoch 40/50
14/14 [==============================] - 1s 48ms/step - loss: 4612957.5000 - mae: 1270.2135 - val_loss: 2112119.2500 - val_mae: 1135.6956
Epoch 41/50
14/14 [==============================] - 1s 50ms/step - loss: 6296884.5000 - mae: 1379.5267 - val_loss: 6763894.0000 - val_mae: 2244.4678
Epoch 42/50
14/14 [==============================] - 1s 47ms/step - loss: 6525176.5000 - mae: 1470.7224 - val_loss: 14936518.0000 - val_mae: 3604.1646
Epoch 43/50
14/14 [==============================] - 1s 49ms/step - loss: 8922662.0000 - mae: 2276.5759 - val_loss: 6590064.5000 - val_mae: 2269.4119
Epoch 44/50
14/14 [==============================] - 1s 49ms/step - loss: 6505676.0000 - mae: 1682.4139 - val_loss: 5616199.5000 - val_mae: 1946.0612
Epoch 45/50
14/14 [==============================] - 1s 47ms/step - loss: 6353706.0000 - mae: 1590.4766 - val_loss: 4748698.0000 - val_mae: 2017.6055
Epoch 46/50
14/14 [==============================] - 1s 48ms/step - loss: 5611137.5000 - mae: 1374.6112 - val_loss: 6451386.0000 - val_mae: 2314.7810
Epoch 47/50
14/14 [==============================] - 1s 49ms/step - loss: 4910579.0000 - mae: 1291.1406 - val_loss: 4041683.0000 - val_mae: 1657.8103
Epoch 48/50
14/14 [==============================] - 1s 47ms/step - loss: 9604768.0000 - mae: 2004.9813 - val_loss: 7787944.0000 - val_mae: 2489.7725
Epoch 49/50
14/14 [==============================] - 1s 48ms/step - loss: 16501388.0000 - mae: 2753.7888 - val_loss: 10653388.0000 - val_mae: 3045.8367
Epoch 50/50
14/14 [==============================] - 1s 46ms/step - loss: 12748582.0000 - mae: 2423.3013 - val_loss: 8023697.5000 - val_mae: 2547.6824


# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


#Plot actual vs. test and validation prediction
shift = 14
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.9)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


# Forecasting 14 days beyond our dataset
n_future=14
forecast = model_lstm_shift.predict(xtrain[-n_future:])


#Plot actual vs. test and validation prediction
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.plot(range(len(y_train)+len(y_test),len(y_train)+len(y_test)+len(forecast)),forecast,color='r',alpha=.8,lw=2,label="Forecast")
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_8=round(mean_MAE[0],4)
MAE_val_8=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_8)
print("The validation MAE of the model is", MAE_val_8)

The MAE of the model is 0.0353
The validation MAE of the model is 0.0493


tf.keras.backend.clear_session()

# Using info New York
info_sel_state = sel_state("California")
info_train_test=train_test("California")

# Scaler
scaler_y = MinMaxScaler(feature_range=(0, 1))

# minmax scaler for y, scaling between 0 and 1
ytrain = info_train_test[1].to_numpy().reshape(-1,1)
ytrain = scaler_y.fit_transform(ytrain)
ytest  = info_train_test[3].to_numpy().reshape(-1,1)
ytest  = scaler_y.fit_transform(ytest)

#Call x variables
xtrain=info_train_test[4]
xtrain.drop(["date", "JHU_cases"], axis=1, inplace=True)
xtest=info_train_test[5].copy()
xtest.drop(["date", "JHU_cases"], axis=1, inplace=True)

# time series generator uses 14 day block to predict the next day 
train_gen = TimeseriesGenerator(xtrain.to_numpy(), ytrain,
                               length=14, sampling_rate=1,  
                                batch_size = 558)

test_gen = TimeseriesGenerator(xtest.to_numpy(), ytest,
                                length=14, sampling_rate=1, 
                               batch_size = 186)

# Below we are creating x and y train and test variable from the generators
x_train, y_train = train_gen[0]
# x_val, y_val = val_gen[0]
x_test, y_test = test_gen[0]

print('x train shape is :', x_train.shape)
print('y train shape is :', y_train.shape)
print('x test shape is :' , x_test.shape)
print('y test shape is :' , y_test.shape)

x train shape is : (544, 14, 498)
y train shape is : (544, 1)
x test shape is : (172, 14, 498)
y test shape is : (172, 1)


# input dimension below
input_dim = x_train.shape[1:]
n_units = 100

#Create model
model_lstm_input = tf.keras.Input(shape=input_dim)
model_hidden=tf.keras.layers.LSTM(units = n_units)(model_lstm_input)
model_lstm_output=tf.keras.layers.Dense(units = 1, activation="linear")(model_hidden)
model_lstm = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm")


#Print the model architecture
print(model_lstm.summary())

#Compile model
model_lstm.compile(optimizer=tf.keras.optimizers.Adam(), loss = 'mse', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_lstm.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_lstm"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 100)               239600    
                                                                 
 dense (Dense)               (None, 1)                 101       
                                                                 
=================================================================
Total params: 239,701
Trainable params: 239,701
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 8s 13ms/step - loss: 0.0242 - mae: 0.0858 - val_loss: 0.0010 - val_mae: 0.0243
Epoch 2/50
489/489 [==============================] - 6s 13ms/step - loss: 0.0050 - mae: 0.0500 - val_loss: 0.0098 - val_mae: 0.0934
Epoch 3/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0049 - mae: 0.0460 - val_loss: 0.0027 - val_mae: 0.0447
Epoch 4/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0036 - mae: 0.0408 - val_loss: 0.0038 - val_mae: 0.0552
Epoch 5/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0031 - mae: 0.0382 - val_loss: 0.0018 - val_mae: 0.0356
Epoch 6/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0030 - mae: 0.0366 - val_loss: 0.0012 - val_mae: 0.0279
Epoch 7/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0034 - mae: 0.0385 - val_loss: 8.8367e-04 - val_mae: 0.0202
Epoch 8/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0028 - mae: 0.0348 - val_loss: 0.0064 - val_mae: 0.0759
Epoch 9/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0028 - mae: 0.0350 - val_loss: 0.0012 - val_mae: 0.0273
Epoch 10/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0024 - mae: 0.0326 - val_loss: 7.4313e-04 - val_mae: 0.0190
Epoch 11/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0022 - mae: 0.0310 - val_loss: 8.7245e-04 - val_mae: 0.0218
Epoch 12/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0022 - mae: 0.0310 - val_loss: 0.0029 - val_mae: 0.0498
Epoch 13/50
489/489 [==============================] - 7s 14ms/step - loss: 0.0029 - mae: 0.0348 - val_loss: 4.6095e-04 - val_mae: 0.0116
Epoch 14/50
489/489 [==============================] - 6s 13ms/step - loss: 0.0020 - mae: 0.0270 - val_loss: 6.3583e-04 - val_mae: 0.0155
Epoch 15/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0017 - mae: 0.0253 - val_loss: 6.7369e-04 - val_mae: 0.0188
Epoch 16/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0014 - mae: 0.0245 - val_loss: 5.4225e-04 - val_mae: 0.0137
Epoch 17/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0015 - mae: 0.0243 - val_loss: 5.4643e-04 - val_mae: 0.0142
Epoch 18/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0014 - mae: 0.0233 - val_loss: 4.9929e-04 - val_mae: 0.0124
Epoch 19/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0013 - mae: 0.0229 - val_loss: 5.1739e-04 - val_mae: 0.0126
Epoch 20/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0015 - mae: 0.0257 - val_loss: 5.7455e-04 - val_mae: 0.0144
Epoch 21/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0017 - mae: 0.0268 - val_loss: 4.9246e-04 - val_mae: 0.0123
Epoch 22/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0011 - mae: 0.0212 - val_loss: 9.1582e-04 - val_mae: 0.0242
Epoch 23/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0011 - mae: 0.0216 - val_loss: 9.7890e-04 - val_mae: 0.0259
Epoch 24/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0011 - mae: 0.0208 - val_loss: 0.0015 - val_mae: 0.0356
Epoch 25/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0013 - mae: 0.0235 - val_loss: 5.9623e-04 - val_mae: 0.0166
Epoch 26/50
489/489 [==============================] - 6s 12ms/step - loss: 9.6111e-04 - mae: 0.0200 - val_loss: 6.7288e-04 - val_mae: 0.0186
Epoch 27/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0012 - mae: 0.0222 - val_loss: 5.5359e-04 - val_mae: 0.0148
Epoch 28/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0011 - mae: 0.0202 - val_loss: 0.0013 - val_mae: 0.0319
Epoch 29/50
489/489 [==============================] - 6s 12ms/step - loss: 0.0012 - mae: 0.0238 - val_loss: 4.8396e-04 - val_mae: 0.0124
Epoch 30/50
489/489 [==============================] - 6s 12ms/step - loss: 9.4039e-04 - mae: 0.0200 - val_loss: 5.7284e-04 - val_mae: 0.0139
Epoch 31/50
489/489 [==============================] - 6s 13ms/step - loss: 9.5435e-04 - mae: 0.0199 - val_loss: 5.1477e-04 - val_mae: 0.0140
Epoch 32/50
489/489 [==============================] - 6s 12ms/step - loss: 8.7772e-04 - mae: 0.0183 - val_loss: 6.2283e-04 - val_mae: 0.0177
Epoch 33/50
489/489 [==============================] - 6s 13ms/step - loss: 7.9906e-04 - mae: 0.0179 - val_loss: 6.0410e-04 - val_mae: 0.0151
Epoch 34/50
489/489 [==============================] - 6s 13ms/step - loss: 8.1934e-04 - mae: 0.0190 - val_loss: 7.8557e-04 - val_mae: 0.0206
Epoch 35/50
489/489 [==============================] - 6s 12ms/step - loss: 7.7926e-04 - mae: 0.0190 - val_loss: 4.6500e-04 - val_mae: 0.0113
Epoch 36/50
489/489 [==============================] - 6s 12ms/step - loss: 9.5465e-04 - mae: 0.0197 - val_loss: 5.8315e-04 - val_mae: 0.0151
Epoch 37/50
489/489 [==============================] - 6s 12ms/step - loss: 6.1789e-04 - mae: 0.0167 - val_loss: 6.0421e-04 - val_mae: 0.0159
Epoch 38/50
489/489 [==============================] - 6s 12ms/step - loss: 9.1427e-04 - mae: 0.0203 - val_loss: 4.8548e-04 - val_mae: 0.0125
Epoch 39/50
489/489 [==============================] - 6s 13ms/step - loss: 6.2995e-04 - mae: 0.0165 - val_loss: 5.6468e-04 - val_mae: 0.0159
Epoch 40/50
489/489 [==============================] - 6s 12ms/step - loss: 7.2378e-04 - mae: 0.0175 - val_loss: 5.7095e-04 - val_mae: 0.0136
Epoch 41/50
489/489 [==============================] - 6s 13ms/step - loss: 5.9768e-04 - mae: 0.0166 - val_loss: 5.2024e-04 - val_mae: 0.0121
Epoch 42/50
489/489 [==============================] - 6s 13ms/step - loss: 8.5951e-04 - mae: 0.0194 - val_loss: 5.4544e-04 - val_mae: 0.0136
Epoch 43/50
489/489 [==============================] - 6s 13ms/step - loss: 5.4284e-04 - mae: 0.0152 - val_loss: 5.4606e-04 - val_mae: 0.0138
Epoch 44/50
489/489 [==============================] - 6s 13ms/step - loss: 7.0342e-04 - mae: 0.0172 - val_loss: 5.5591e-04 - val_mae: 0.0133
Epoch 45/50
489/489 [==============================] - 6s 13ms/step - loss: 6.6437e-04 - mae: 0.0174 - val_loss: 5.5653e-04 - val_mae: 0.0146
Epoch 46/50
489/489 [==============================] - 6s 13ms/step - loss: 5.7819e-04 - mae: 0.0161 - val_loss: 5.0407e-04 - val_mae: 0.0117
Epoch 47/50
489/489 [==============================] - 7s 14ms/step - loss: 5.2939e-04 - mae: 0.0152 - val_loss: 5.5468e-04 - val_mae: 0.0129
Epoch 48/50
489/489 [==============================] - 6s 12ms/step - loss: 6.8378e-04 - mae: 0.0172 - val_loss: 8.4082e-04 - val_mae: 0.0219
Epoch 49/50
489/489 [==============================] - 6s 13ms/step - loss: 4.9727e-04 - mae: 0.0155 - val_loss: 5.7204e-04 - val_mae: 0.0143
Epoch 50/50
489/489 [==============================] - 6s 13ms/step - loss: 4.0654e-04 - mae: 0.0136 - val_loss: 5.0659e-04 - val_mae: 0.0120


# Plotting the MSE and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of one-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of one-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')

axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
lstm_prediction = scaler_y.inverse_transform(model_lstm.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), lstm_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (one-layer LSTM model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_9=round(mean_MAE[0],4)
MAE_val_9=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_9)
print("The validation MAE of the model is", MAE_val_9)

The MAE of the model is 0.0255
The validation MAE of the model is 0.0221


### Multilayer LSTM model
#Design following: Iberoamerican Journal of Medicine
tf.keras.backend.clear_session()

# input dimension below
input_dim = x_train.shape[1:]
n_units= 50

#Create model
model_lstm_input = tf.keras.Input(shape=input_dim)
model_hidden1=tf.keras.layers.LSTM(units = n_units, return_sequences=True)(model_lstm_input)
model_dropout1=tf.keras.layers.Dropout(rate=0.2)(model_hidden1)
model_hidden2=tf.keras.layers.LSTM(units = n_units, return_sequences=True)(model_dropout1)
model_dropout2=tf.keras.layers.Dropout(rate=0.2)(model_hidden2)
model_hidden3=tf.keras.layers.LSTM(units = n_units, return_sequences=False)(model_dropout2)
model_dropout3=tf.keras.layers.Dropout(rate=0.2)(model_hidden3)
model_lstm_output=tf.keras.layers.Dense(units = 1, activation="linear")(model_dropout3)
model_lstm = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm")

#Print the model architecture
print(model_lstm.summary())

#Compile model
model_lstm.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Convert x to tensor
x_train_tf = tf.convert_to_tensor(x_train, np.float32)

#Fit model
history = model_lstm.fit(x_train_tf, y_train, epochs = 50, validation_split=0.1, batch_size = 1, verbose=1)

Model: "model_lstm"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 50)            109800    
                                                                 
 dropout (Dropout)           (None, 14, 50)            0         
                                                                 
 lstm_1 (LSTM)               (None, 14, 50)            20200     
                                                                 
 dropout_1 (Dropout)         (None, 14, 50)            0         
                                                                 
 lstm_2 (LSTM)               (None, 50)                20200     
                                                                 
 dropout_2 (Dropout)         (None, 50)                0         
                                                                 
 dense (Dense)               (None, 1)                 51        
                                                                 
=================================================================
Total params: 150,251
Trainable params: 150,251
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
489/489 [==============================] - 15s 19ms/step - loss: 0.0199 - mae: 0.0914 - val_loss: 6.4720e-04 - val_mae: 0.0210
Epoch 2/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0081 - mae: 0.0547 - val_loss: 0.0063 - val_mae: 0.0771
Epoch 3/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0052 - mae: 0.0450 - val_loss: 3.9298e-04 - val_mae: 0.0107
Epoch 4/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0064 - mae: 0.0493 - val_loss: 0.0034 - val_mae: 0.0558
Epoch 5/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0046 - mae: 0.0411 - val_loss: 0.0014 - val_mae: 0.0335
Epoch 6/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0045 - mae: 0.0416 - val_loss: 7.6339e-04 - val_mae: 0.0238
Epoch 7/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0039 - mae: 0.0345 - val_loss: 3.9490e-04 - val_mae: 0.0121
Epoch 8/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0054 - mae: 0.0439 - val_loss: 0.0015 - val_mae: 0.0365
Epoch 9/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0043 - mae: 0.0393 - val_loss: 0.0017 - val_mae: 0.0385
Epoch 10/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0037 - mae: 0.0362 - val_loss: 4.0640e-04 - val_mae: 0.0121
Epoch 11/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0036 - mae: 0.0348 - val_loss: 8.9209e-04 - val_mae: 0.0266
Epoch 12/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0038 - mae: 0.0366 - val_loss: 5.6357e-04 - val_mae: 0.0185
Epoch 13/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0039 - mae: 0.0359 - val_loss: 3.8485e-04 - val_mae: 0.0106
Epoch 14/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0036 - mae: 0.0349 - val_loss: 5.5303e-04 - val_mae: 0.0178
Epoch 15/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0040 - mae: 0.0361 - val_loss: 4.1175e-04 - val_mae: 0.0125
Epoch 16/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0033 - mae: 0.0339 - val_loss: 4.2312e-04 - val_mae: 0.0115
Epoch 17/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0034 - mae: 0.0318 - val_loss: 5.6858e-04 - val_mae: 0.0185
Epoch 18/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0036 - mae: 0.0349 - val_loss: 9.5436e-04 - val_mae: 0.0275
Epoch 19/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0039 - mae: 0.0389 - val_loss: 0.0014 - val_mae: 0.0349
Epoch 20/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0040 - mae: 0.0368 - val_loss: 0.0029 - val_mae: 0.0516
Epoch 21/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0032 - mae: 0.0325 - val_loss: 0.0038 - val_mae: 0.0589
Epoch 22/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0024 - mae: 0.0314 - val_loss: 5.1646e-04 - val_mae: 0.0168
Epoch 23/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0035 - mae: 0.0351 - val_loss: 3.9534e-04 - val_mae: 0.0106
Epoch 24/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0031 - mae: 0.0341 - val_loss: 6.8110e-04 - val_mae: 0.0219
Epoch 25/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0037 - mae: 0.0367 - val_loss: 4.9154e-04 - val_mae: 0.0154
Epoch 26/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0028 - mae: 0.0303 - val_loss: 0.0027 - val_mae: 0.0493
Epoch 27/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0029 - mae: 0.0320 - val_loss: 4.4094e-04 - val_mae: 0.0134
Epoch 28/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0030 - mae: 0.0331 - val_loss: 4.0304e-04 - val_mae: 0.0113
Epoch 29/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0024 - mae: 0.0285 - val_loss: 5.7885e-04 - val_mae: 0.0189
Epoch 30/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0032 - mae: 0.0340 - val_loss: 6.9082e-04 - val_mae: 0.0214
Epoch 31/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0024 - mae: 0.0291 - val_loss: 8.4675e-04 - val_mae: 0.0257
Epoch 32/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0026 - mae: 0.0308 - val_loss: 5.6727e-04 - val_mae: 0.0184
Epoch 33/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0032 - mae: 0.0320 - val_loss: 6.8826e-04 - val_mae: 0.0219
Epoch 34/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0028 - mae: 0.0312 - val_loss: 6.6016e-04 - val_mae: 0.0207
Epoch 35/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0026 - mae: 0.0299 - val_loss: 5.3233e-04 - val_mae: 0.0169
Epoch 36/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0025 - mae: 0.0288 - val_loss: 9.1059e-04 - val_mae: 0.0269
Epoch 37/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0029 - mae: 0.0289 - val_loss: 4.2710e-04 - val_mae: 0.0124
Epoch 38/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0023 - mae: 0.0280 - val_loss: 8.7043e-04 - val_mae: 0.0261
Epoch 39/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0027 - mae: 0.0301 - val_loss: 4.8235e-04 - val_mae: 0.0128
Epoch 40/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0028 - mae: 0.0300 - val_loss: 6.9948e-04 - val_mae: 0.0222
Epoch 41/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0028 - mae: 0.0303 - val_loss: 5.2715e-04 - val_mae: 0.0170
Epoch 42/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0020 - mae: 0.0283 - val_loss: 4.7490e-04 - val_mae: 0.0144
Epoch 43/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0027 - mae: 0.0296 - val_loss: 0.0012 - val_mae: 0.0325
Epoch 44/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0022 - mae: 0.0298 - val_loss: 4.2577e-04 - val_mae: 0.0107
Epoch 45/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0021 - mae: 0.0268 - val_loss: 7.0891e-04 - val_mae: 0.0200
Epoch 46/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0026 - mae: 0.0300 - val_loss: 5.5670e-04 - val_mae: 0.0180
Epoch 47/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0028 - mae: 0.0296 - val_loss: 0.0021 - val_mae: 0.0436
Epoch 48/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0022 - mae: 0.0287 - val_loss: 4.5472e-04 - val_mae: 0.0142
Epoch 49/50
489/489 [==============================] - 8s 16ms/step - loss: 0.0023 - mae: 0.0288 - val_loss: 4.2030e-04 - val_mae: 0.0104
Epoch 50/50
489/489 [==============================] - 8s 17ms/step - loss: 0.0020 - mae: 0.0264 - val_loss: 4.0863e-04 - val_mae: 0.0100


# Plotting the MSE and MAE graphs
# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer LSTM model for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), history.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


# Predicting the y variable

# we are using inverse transform below to convert the minmax scaled y to the original cases
lstm_prediction = scaler_y.inverse_transform(model_lstm.predict(x_test))


#Transform date to string    
xtest_df=info_train_test[5]
dates=xtest_df["date"]
list_dates=[]
for i in dates:
    date=i.strftime("%d/%m/%Y")
    list_dates.append(date)
    
#Plot
plt.figure(figsize=(12,6))
plt.plot(range(len(y_test)), lstm_prediction, label="Predicted")
plt.plot(range(len(y_test)),list(scaler_y.inverse_transform(y_test)), label="Actual")
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM model) for {file_names[info_sel_state[0]]} using test data")
plt.xticks(range(len(ytest))[::4], list_dates[::4], rotation ="vertical", fontsize= 8)
plt.xlabel("Date")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_10=round(mean_MAE[0],4)
MAE_val_10=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_10)
print("The validation MAE of the model is", MAE_val_10)

The MAE of the model is 0.0349
The validation MAE of the model is 0.0237


#Using info New York
info_sel_state = sel_state("California")
info_train_test=train_test("California")

#x and y train and test variables
x_train=info_train_test[4]
x_train.drop(["date"], axis=1, inplace=True)
y_train=info_train_test[1]
x_test=info_train_test[5]
x_test.drop(["date"], axis=1, inplace=True)
y_test=info_train_test[3]

# Datasets shape
print("\nThe shape of xtrain is", x_train.shape)
print("The shape of ytrain is", y_train.shape)

print("\nThe shape of xtest is", x_test.shape)
print("The shape of ytest is", y_test.shape)

The shape of xtrain is (558, 499)
The shape of ytrain is (558,)

The shape of xtest is (186, 499)
The shape of ytest is (186,)


# Function to create datasets for LSTM as time steps
def format_series(n_future,n_past,df):
  # Creates series of multi-step on the dataset based on n_past periods
  # and adds the corresponding predicted values based on n_future periods
    trainX=[]
    trainY=[]
    for i in range(n_past,len(df)-n_future+1):
        trainX.append(df.iloc[i-n_past:i,1:df.shape[1]])#All features except y
        trainY.append(df.iloc[i+n_future-1:i+n_future,0])#y

    return np.array(trainX),np.array(trainY)


# Generating Time Step Series
xtrain,ytrain = format_series(1,14,x_train) 
xtest,ytest = format_series(1,14,x_test) 

# Checking the test and train data shapes
print("\nThe shape of xtrain is", xtrain.shape)
print("The shape of ytrain is", ytrain.shape)

print("\nThe shape of xtest is", xtest.shape)
print("The shape of ytest is", ytest.shape)

The shape of xtrain is (544, 14, 498)
The shape of ytrain is (544, 1)

The shape of xtest is (172, 14, 498)
The shape of ytest is (172, 1)


tf.keras.backend.clear_session()

n_units=64

# Model architecture with 2 LSTM layers
model_lstm_input = tf.keras.Input(shape=(xtrain.shape[1], xtrain.shape[2]))
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_lstm_input)
model_hidden=tf.keras.layers.LSTM(units = n_units//2, activation='relu', return_sequences=False)(model_hidden)
model_hidden=tf.keras.layers.Dropout(0.2)(model_hidden)
model_lstm_output=tf.keras.layers.Dense(units = 1)(model_hidden)

model_lstm_shift = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm_time_steps")

#Print the model architecture
print(model_lstm_shift.summary())

#Compile model
model_lstm_shift.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Fit model
model_lstm_shift_results=model_lstm_shift.fit(xtrain, ytrain, 
                                               validation_split=0.2, 
                                               epochs = 50,
                                               batch_size = 32)

#Predict with model
model_lstm_shift_predict=model_lstm_shift.predict(xtest)

Model: "model_lstm_time_steps"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 64)            144128    
                                                                 
 lstm_1 (LSTM)               (None, 32)                12416     
                                                                 
 dropout (Dropout)           (None, 32)                0         
                                                                 
 dense (Dense)               (None, 1)                 33        
                                                                 
=================================================================
Total params: 156,577
Trainable params: 156,577
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/50
14/14 [==============================] - 4s 101ms/step - loss: 211965520.0000 - mae: 8498.7295 - val_loss: 3273468.5000 - val_mae: 1434.0105
Epoch 2/50
14/14 [==============================] - 0s 31ms/step - loss: 181791840.0000 - mae: 7388.6646 - val_loss: 30547176.0000 - val_mae: 5329.3003
Epoch 3/50
14/14 [==============================] - 0s 30ms/step - loss: 142289808.0000 - mae: 7866.9229 - val_loss: 46882304.0000 - val_mae: 6680.9351
Epoch 4/50
14/14 [==============================] - 0s 31ms/step - loss: 134673168.0000 - mae: 7133.2437 - val_loss: 52617780.0000 - val_mae: 7106.0850
Epoch 5/50
14/14 [==============================] - 0s 33ms/step - loss: 126756376.0000 - mae: 7416.6733 - val_loss: 65704912.0000 - val_mae: 7975.9336
Epoch 6/50
14/14 [==============================] - 0s 31ms/step - loss: 69502040.0000 - mae: 5036.1968 - val_loss: 15938108.0000 - val_mae: 3839.4763
Epoch 7/50
14/14 [==============================] - 0s 32ms/step - loss: 35582600.0000 - mae: 3916.6445 - val_loss: 3540142.5000 - val_mae: 1704.6936
Epoch 8/50
14/14 [==============================] - 0s 31ms/step - loss: 90303768.0000 - mae: 5564.8311 - val_loss: 1558467.2500 - val_mae: 782.2410
Epoch 9/50
14/14 [==============================] - 0s 32ms/step - loss: 86665392.0000 - mae: 5239.8696 - val_loss: 2893912.0000 - val_mae: 1113.5511
Epoch 10/50
14/14 [==============================] - 0s 33ms/step - loss: 63810940.0000 - mae: 4602.9082 - val_loss: 1358195.3750 - val_mae: 693.4846
Epoch 11/50
14/14 [==============================] - 0s 32ms/step - loss: 42374216.0000 - mae: 3933.2227 - val_loss: 3261521.5000 - val_mae: 1610.9432
Epoch 12/50
14/14 [==============================] - 0s 32ms/step - loss: 189152208.0000 - mae: 11158.7637 - val_loss: 53090464.0000 - val_mae: 7171.8066
Epoch 13/50
14/14 [==============================] - 0s 32ms/step - loss: 106676704.0000 - mae: 5913.0352 - val_loss: 1287265.5000 - val_mae: 740.5115
Epoch 14/50
14/14 [==============================] - 0s 32ms/step - loss: 120201680.0000 - mae: 6075.6958 - val_loss: 26331160.0000 - val_mae: 4986.2031
Epoch 15/50
14/14 [==============================] - 0s 32ms/step - loss: 106026856.0000 - mae: 6290.9160 - val_loss: 46089272.0000 - val_mae: 6662.6294
Epoch 16/50
14/14 [==============================] - 0s 32ms/step - loss: 100596432.0000 - mae: 6607.5190 - val_loss: 51968788.0000 - val_mae: 7091.3540
Epoch 17/50
14/14 [==============================] - 0s 29ms/step - loss: 75256664.0000 - mae: 5399.1514 - val_loss: 21754054.0000 - val_mae: 3889.6436
Epoch 18/50
14/14 [==============================] - 0s 33ms/step - loss: 71420848.0000 - mae: 5454.8418 - val_loss: 14686354.0000 - val_mae: 3266.7651
Epoch 19/50
14/14 [==============================] - 0s 31ms/step - loss: 56937136.0000 - mae: 3941.4146 - val_loss: 7949147.0000 - val_mae: 2160.6599
Epoch 20/50
14/14 [==============================] - 0s 34ms/step - loss: 44117388.0000 - mae: 3660.1829 - val_loss: 2335649.7500 - val_mae: 1077.7430
Epoch 21/50
14/14 [==============================] - 0s 33ms/step - loss: 23467996.0000 - mae: 2953.9502 - val_loss: 2447765.0000 - val_mae: 1100.4504
Epoch 22/50
14/14 [==============================] - 0s 31ms/step - loss: 30035370.0000 - mae: 3217.8740 - val_loss: 7498751.5000 - val_mae: 2234.8191
Epoch 23/50
14/14 [==============================] - 0s 32ms/step - loss: 27539032.0000 - mae: 2915.3477 - val_loss: 4217333.5000 - val_mae: 1643.8883
Epoch 24/50
14/14 [==============================] - 0s 32ms/step - loss: 25933814.0000 - mae: 2969.0520 - val_loss: 2760289.0000 - val_mae: 1160.8108
Epoch 25/50
14/14 [==============================] - 0s 32ms/step - loss: 26840592.0000 - mae: 2941.1904 - val_loss: 4074686.2500 - val_mae: 1537.7198
Epoch 26/50
14/14 [==============================] - 0s 31ms/step - loss: 23761842.0000 - mae: 2786.1062 - val_loss: 3727128.0000 - val_mae: 1432.5961
Epoch 27/50
14/14 [==============================] - 0s 32ms/step - loss: 27881032.0000 - mae: 3652.7236 - val_loss: 13375986.0000 - val_mae: 3484.3940
Epoch 28/50
14/14 [==============================] - 0s 33ms/step - loss: 24536220.0000 - mae: 3172.0112 - val_loss: 2735400.0000 - val_mae: 1352.9434
Epoch 29/50
14/14 [==============================] - 0s 35ms/step - loss: 103220744.0000 - mae: 6031.9404 - val_loss: 3086850.2500 - val_mae: 1369.8469
Epoch 30/50
14/14 [==============================] - 1s 36ms/step - loss: 62254900.0000 - mae: 4905.7773 - val_loss: 4037229.7500 - val_mae: 1697.7083
Epoch 31/50
14/14 [==============================] - 0s 32ms/step - loss: 151210784.0000 - mae: 7002.9478 - val_loss: 6006857.0000 - val_mae: 2227.8179
Epoch 32/50
14/14 [==============================] - 0s 31ms/step - loss: 126095808.0000 - mae: 6573.0713 - val_loss: 66913372.0000 - val_mae: 8053.4209
Epoch 33/50
14/14 [==============================] - 0s 32ms/step - loss: 70847520.0000 - mae: 5824.7715 - val_loss: 36431924.0000 - val_mae: 5884.0815
Epoch 34/50
14/14 [==============================] - 0s 31ms/step - loss: 33594176.0000 - mae: 4221.6182 - val_loss: 3612692.5000 - val_mae: 1564.3646
Epoch 35/50
14/14 [==============================] - 0s 31ms/step - loss: 32311990.0000 - mae: 3849.3606 - val_loss: 1553731.8750 - val_mae: 875.6566
Epoch 36/50
14/14 [==============================] - 0s 31ms/step - loss: 26592736.0000 - mae: 3325.4312 - val_loss: 8159780.5000 - val_mae: 2691.9521
Epoch 37/50
14/14 [==============================] - 0s 32ms/step - loss: 24768300.0000 - mae: 3255.9922 - val_loss: 3874033.0000 - val_mae: 1696.5494
Epoch 38/50
14/14 [==============================] - 0s 32ms/step - loss: 24107874.0000 - mae: 3265.0774 - val_loss: 6569189.5000 - val_mae: 2373.1074
Epoch 39/50
14/14 [==============================] - 0s 33ms/step - loss: 24603514.0000 - mae: 3211.7834 - val_loss: 3945404.0000 - val_mae: 1747.2844
Epoch 40/50
14/14 [==============================] - 0s 32ms/step - loss: 21820846.0000 - mae: 3013.8655 - val_loss: 4104106.7500 - val_mae: 1806.1414
Epoch 41/50
14/14 [==============================] - 0s 32ms/step - loss: 20923104.0000 - mae: 2963.4680 - val_loss: 2539174.2500 - val_mae: 1290.9636
Epoch 42/50
14/14 [==============================] - 0s 33ms/step - loss: 33846612.0000 - mae: 4031.4465 - val_loss: 4425227.0000 - val_mae: 1849.7350
Epoch 43/50
14/14 [==============================] - 0s 35ms/step - loss: 45077320.0000 - mae: 4643.5820 - val_loss: 1929012.8750 - val_mae: 1058.1930
Epoch 44/50
14/14 [==============================] - 0s 31ms/step - loss: 31429124.0000 - mae: 3890.6455 - val_loss: 6678005.5000 - val_mae: 2278.4661
Epoch 45/50
14/14 [==============================] - 0s 31ms/step - loss: 25508872.0000 - mae: 3629.1709 - val_loss: 7739110.0000 - val_mae: 2525.3257
Epoch 46/50
14/14 [==============================] - 0s 31ms/step - loss: 27984904.0000 - mae: 3352.6294 - val_loss: 3625606.7500 - val_mae: 1648.5385
Epoch 47/50
14/14 [==============================] - 0s 33ms/step - loss: 25264998.0000 - mae: 3408.8630 - val_loss: 5586283.0000 - val_mae: 2151.0068
Epoch 48/50
14/14 [==============================] - 0s 32ms/step - loss: 26763844.0000 - mae: 3248.0813 - val_loss: 1291578.8750 - val_mae: 802.8881
Epoch 49/50
14/14 [==============================] - 0s 32ms/step - loss: 22768166.0000 - mae: 3051.8604 - val_loss: 2150007.0000 - val_mae: 1233.3048
Epoch 50/50
14/14 [==============================] - 0s 33ms/step - loss: 20158450.0000 - mae: 2774.0388 - val_loss: 3659237.5000 - val_mae: 1691.2209


# Plot training results
epochs=50
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of two-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of two-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')
axs.legend()

plt.show()


#Plot actual vs. test and validation prediction
shift = 14
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.title(f"Predicted versus actual JHU_cases (two-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


# Forecasting 14 days beyond our dataset
n_future=14
forecast = model_lstm_shift.predict(xtrain[-n_future:])


#Plot actual vs. test and validation prediction
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.plot(range(len(y_train)+len(y_test),len(y_train)+len(y_test)+len(forecast)),forecast,color='r',alpha=.8,lw=2,label="Forecast")
plt.title(f"Predicted versus actual JHU_cases (two-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_11=round(mean_MAE[0],4)
MAE_val_11=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_11)
print("The validation MAE of the model is", MAE_val_11)

The MAE of the model is 0.0349
The validation MAE of the model is 0.0237


#Using info New York
info_sel_state = sel_state("California")
info_train_test=train_test("California")

#x and y train and test variables
x_train=info_train_test[4]
x_train.drop(["date"], axis=1, inplace=True)
y_train=info_train_test[1]
x_test=info_train_test[5]
x_test.drop(["date"], axis=1, inplace=True)
y_test=info_train_test[3]

# Datasets shape
print("\nThe shape of xtrain is", x_train.shape)
print("The shape of ytrain is", y_train.shape)

print("\nThe shape of xtest is", x_test.shape)
print("The shape of ytest is", y_test.shape)

The shape of xtrain is (558, 499)
The shape of ytrain is (558,)

The shape of xtest is (186, 499)
The shape of ytest is (186,)


# Generating Time Step Series
xtrain,ytrain = format_series(1,14,x_train) 
xtest,ytest = format_series(1,14,x_test) 

# Checking the test and train data shapes
print("\nThe shape of xtrain is", xtrain.shape)
print("The shape of ytrain is", ytrain.shape)

print("\nThe shape of xtest is", xtest.shape)
print("The shape of ytest is", ytest.shape)

The shape of xtrain is (544, 14, 498)
The shape of ytrain is (544, 1)

The shape of xtest is (172, 14, 498)
The shape of ytest is (172, 1)


tf.keras.backend.clear_session()

n_units=64

# Model architecture with 4 LSTM layers
model_lstm_input = tf.keras.Input(shape=(xtrain.shape[1], xtrain.shape[2]))
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_lstm_input)
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_hidden)
model_hidden=tf.keras.layers.LSTM(units = n_units,    activation='relu', return_sequences=True)(model_hidden)
model_hidden=tf.keras.layers.LSTM(units = n_units//2, activation='relu', return_sequences=False)(model_hidden)
model_hidden=tf.keras.layers.Dropout(0.2)(model_hidden)
model_lstm_output=tf.keras.layers.Dense(units = 1)(model_hidden)

model_lstm_shift = tf.keras.Model(inputs=model_lstm_input, outputs=model_lstm_output, name="model_lstm_time_steps")

#Print the model architecture
print(model_lstm_shift.summary())

#Compile model
model_lstm_shift.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics=['mae'])

#Fit model
model_lstm_shift_results=model_lstm_shift.fit(xtrain, ytrain, 
                                               validation_split=0.2, 
                                               epochs = 20,
                                               batch_size = 32)

#Predict with model
model_lstm_shift_predict=model_lstm_shift.predict(xtest)

Model: "model_lstm_time_steps"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 14, 498)]         0         
                                                                 
 lstm (LSTM)                 (None, 14, 64)            144128    
                                                                 
 lstm_1 (LSTM)               (None, 14, 64)            33024     
                                                                 
 lstm_2 (LSTM)               (None, 14, 64)            33024     
                                                                 
 lstm_3 (LSTM)               (None, 32)                12416     
                                                                 
 dropout (Dropout)           (None, 32)                0         
                                                                 
 dense (Dense)               (None, 1)                 33        
                                                                 
=================================================================
Total params: 222,625
Trainable params: 222,625
Non-trainable params: 0
_________________________________________________________________
None
Epoch 1/20
14/14 [==============================] - 6s 99ms/step - loss: 210854784.0000 - mae: 8456.4004 - val_loss: 1626989.7500 - val_mae: 824.3406
Epoch 2/20
14/14 [==============================] - 1s 47ms/step - loss: 166517856.0000 - mae: 7246.9727 - val_loss: 40408248.0000 - val_mae: 6209.0830
Epoch 3/20
14/14 [==============================] - 1s 49ms/step - loss: 136117760.0000 - mae: 7788.8120 - val_loss: 91118264.0000 - val_mae: 9415.2812
Epoch 4/20
14/14 [==============================] - 1s 49ms/step - loss: 124378032.0000 - mae: 7829.9971 - val_loss: 69668576.0000 - val_mae: 8204.8271
Epoch 5/20
14/14 [==============================] - 1s 48ms/step - loss: 123590880.0000 - mae: 7596.1841 - val_loss: 136285008.0000 - val_mae: 11590.5186
Epoch 6/20
14/14 [==============================] - 1s 48ms/step - loss: 116589896.0000 - mae: 7600.4536 - val_loss: 70615160.0000 - val_mae: 8280.9229
Epoch 7/20
14/14 [==============================] - 1s 50ms/step - loss: 97223928.0000 - mae: 6433.7227 - val_loss: 59854392.0000 - val_mae: 7606.1362
Epoch 8/20
14/14 [==============================] - 1s 46ms/step - loss: 81823400.0000 - mae: 5415.7090 - val_loss: 62635788.0000 - val_mae: 7798.8047
Epoch 9/20
14/14 [==============================] - 1s 55ms/step - loss: 65767900.0000 - mae: 5412.9243 - val_loss: 41635440.0000 - val_mae: 6194.2935
Epoch 10/20
14/14 [==============================] - 1s 92ms/step - loss: 53672460.0000 - mae: 4855.0679 - val_loss: 34933376.0000 - val_mae: 5778.6099
Epoch 11/20
14/14 [==============================] - 1s 91ms/step - loss: 68724688.0000 - mae: 5481.1489 - val_loss: 3864430.0000 - val_mae: 1815.8246
Epoch 12/20
14/14 [==============================] - 1s 102ms/step - loss: 60192428.0000 - mae: 4439.5044 - val_loss: 29458972.0000 - val_mae: 5297.2212
Epoch 13/20
14/14 [==============================] - 1s 86ms/step - loss: 48400068.0000 - mae: 4248.7925 - val_loss: 4137835.5000 - val_mae: 1897.6862
Epoch 14/20
14/14 [==============================] - 1s 90ms/step - loss: 41948700.0000 - mae: 3560.7749 - val_loss: 5535162.5000 - val_mae: 2220.7200
Epoch 15/20
14/14 [==============================] - 1s 106ms/step - loss: 31208060.0000 - mae: 3005.7080 - val_loss: 2950685.7500 - val_mae: 1422.2139
Epoch 16/20
14/14 [==============================] - 1s 104ms/step - loss: 24787758.0000 - mae: 2893.1052 - val_loss: 1969769.5000 - val_mae: 1103.8224
Epoch 17/20
14/14 [==============================] - 2s 108ms/step - loss: 37147716.0000 - mae: 3289.9250 - val_loss: 1402637.3750 - val_mae: 795.3813
Epoch 18/20
14/14 [==============================] - 2s 111ms/step - loss: 35615892.0000 - mae: 3279.9409 - val_loss: 14235012.0000 - val_mae: 3608.4495
Epoch 19/20
14/14 [==============================] - 1s 94ms/step - loss: 31004794.0000 - mae: 3404.5710 - val_loss: 8967538.0000 - val_mae: 2756.7134
Epoch 20/20
14/14 [==============================] - 1s 100ms/step - loss: 44776784.0000 - mae: 3630.9194 - val_loss: 1412234.0000 - val_mae: 831.3953


# Plot training results
epochs=20
fig = plt.figure(figsize=(20,5))
axs = fig.add_subplot(1,3,1)
axs.set_title(f'Loss of multi-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["loss","val_loss"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('Loss')
    axs.set_xlabel('epochs')
axs.legend()

axs = fig.add_subplot(1,3,2)
axs.set_title(f'MAE of multi-layer LSTM with time steps for {file_names[info_sel_state[0]]}')
# Plot all metrics
for metric in ["mae","val_mae"]:
    axs.plot(np.arange(0, epochs), model_lstm_shift_results.history[metric], label=metric)
    axs.set_ylabel('MAE')
    axs.set_xlabel('epochs')

axs.legend()

plt.show()


#Plot actual vs. test and validation prediction
shift = 14
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.9)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


# Forecasting 14 days beyond our dataset
n_future=14
forecast = model_lstm_shift.predict(xtrain[-n_future:])


#Plot actual vs. test and validation prediction
plt.figure(figsize=(12,6))
plt.plot(range(len(y_train)),y_train, label="Actual",color='orange',alpha=.8)
plt.plot(range(len(y_train),len(y_train)+len(y_test)),y_test,color='orange',alpha=.8)
plt.plot(range(shift,len(ytrain)+shift), model_lstm_shift.predict(xtrain), label="Validation",color='b',lw=2)
plt.plot(range(len(ytrain)+shift,len(ytrain)+len(ytest)+shift), model_lstm_shift.predict(xtest), label="Predicted",color='g',lw=2)
plt.plot(range(len(y_train)+len(y_test),len(y_train)+len(y_test)+len(forecast)),forecast,color='r',alpha=.8,lw=2,label="Forecast")
plt.title(f"Predicted versus actual JHU_cases (multi-layer LSTM with time steps) for {file_names[info_sel_state[0]]}")
plt.xlabel("Time (in days)")
plt.ylabel("JHU_cases")
plt.legend()
plt.show()


#Calculate mean MAE
mean_MAE=[]
for metric in ["mae","val_mae"]:
    mean_MAE.append(np.mean(history.history[metric]))

MAE_12=round(mean_MAE[0],4)
MAE_val_12=round(mean_MAE[1],4)

print("The MAE of the model is", MAE_12)
print("The validation MAE of the model is", MAE_val_12)

The MAE of the model is 0.0349
The validation MAE of the model is 0.0237


#Summary table of performance of models

#For validation data
pd.set_option('display.float_format', lambda x: '%.4f' % x)

table_names=["ARMA", "One-layer RNN", "Multi-layer RNN", "One-layer LSTM", "Multi-layer LSTM",
             "Two-layer LSTM w/ time steps", "Multi-layer LSTM w/ time steps"]
table_values_ny=[MAE_0_ny, MAE_val_1,MAE_val_2,MAE_val_5,MAE_val_6,MAE_val_7,MAE_val_8]
table_values_cal=[MAE_0_cal, MAE_val_3, MAE_val_4,MAE_val_9,MAE_val_10,MAE_val_11,MAE_val_12]

table_mae = pd.DataFrame(list(zip(table_values_ny, table_values_cal)),
               columns =['New York', 'California'], index=table_names)

table_mae.index.name = 'MAE values'

print("The validation mean absolute error (MAE) of the different models included in this notebook are below:\n")
print(table_mae)

#For train data

table_values_ny=[MAE_0_ny, MAE_1,MAE_2,MAE_5,MAE_6,MAE_7,MAE_8]
table_values_cal=[MAE_0_cal, MAE_3, MAE_4,MAE_9,MAE_10,MAE_11,MAE_12]

table_mae = pd.DataFrame(list(zip(table_values_ny, table_values_cal)),
               columns =['New York', 'California'], index=table_names)

table_mae.index.name = 'MAE values'

print("\n\nThe train mean absolute error (MAE) of the different models included in this notebook are below:\n")
print(table_mae)

The validation mean absolute error (MAE) of the different models included in this notebook are below:

                                New York  California
MAE values                                          
ARMA                              0.1295      0.0658
One-layer RNN                     0.0602      0.0430
Multi-layer RNN                   0.0817      0.0544
One-layer LSTM                    0.0498      0.0221
Multi-layer LSTM                  0.0493      0.0237
Two-layer LSTM w/ time steps      0.0493      0.0237
Multi-layer LSTM w/ time steps    0.0493      0.0237


The train mean absolute error (MAE) of the different models included in this notebook are below:

                                New York  California
MAE values                                          
ARMA                              0.1295      0.0658
One-layer RNN                     0.0408      0.0446
Multi-layer RNN                   0.0601      0.0715
One-layer LSTM                    0.0216      0.0255
Multi-layer LSTM                  0.0353      0.0349
Two-layer LSTM w/ time steps      0.0353      0.0349
Multi-layer LSTM w/ time steps    0.0353      0.0349

	date	JHU_cases	JHU_deaths	JHU_hospitalizations	up2date	gt_chest pain	gt_nose bleed	gt_fever	gt2_Abdominal obesity	gt2_Abdominal pain	gt2_Acne	gt2_Actinic keratosis	gt2_Acute bronchitis	gt2_Adrenal crisis	gt2_Ageusia	gt2_Alcoholism	gt2_Allergic conjunctivitis	gt2_Allergy	gt2_Amblyopia	gt2_Amenorrhea	gt2_Amnesia	gt2_Anal fissure	gt2_Anaphylaxis	gt2_Anemia	gt2_Angina pectoris	gt2_Angioedema	gt2_Angular cheilitis	gt2_Anosmia	gt2_Anxiety	gt2_Aphasia	gt2_Aphonia	gt2_Apnea	gt2_Arthralgia	gt2_Arthritis	gt2_Ascites	gt2_Asperger syndrome	gt2_Asphyxia	gt2_Asthma	gt2_Astigmatism	gt2_Ataxia	gt2_Atheroma	gt2_Attention deficit hyperactivity disorder	gt2_Auditory hallucination	gt2_Autoimmune disease	gt2_Avoidant personality disorder	gt2_Back pain	gt2_Bacterial vaginosis	gt2_Balance disorder	gt2_Beau's lines	gt2_Bell's palsy	gt2_Biliary colic	gt2_Binge eating	gt2_Bleeding	gt2_Bleeding on probing	gt2_Blepharospasm	gt2_Bloating	gt2_Blood in stool	gt2_Blurred vision	gt2_Blushing	gt2_Boil	gt2_Bone fracture	gt2_Bone tumor	gt2_Bowel obstruction	gt2_Bradycardia	gt2_Braxton Hicks contractions	gt2_Breakthrough bleeding	gt2_Breast pain	gt2_Bronchitis	gt2_Bruise	gt2_Bruxism	gt2_Bunion	gt2_Burn	gt2_Burning Chest Pain	gt2_Burning mouth syndrome	gt2_Candidiasis	gt2_Canker sore	gt2_Cardiac arrest	gt2_Carpal tunnel syndrome	gt2_Cataplexy	gt2_Cataract	gt2_Chancre	gt2_Cheilitis	gt2_Chest pain	gt2_Chills	gt2_Chorea	gt2_Chronic pain	gt2_Cirrhosis	gt2_Cleft lip and cleft palate	gt2_Clouding of consciousness	gt2_Cluster headache	gt2_Colitis	gt2_Coma	gt2_Common cold	gt2_Compulsive behavior	gt2_Compulsive hoarding	gt2_Confusion	gt2_Congenital heart defect	gt2_Conjunctivitis	gt2_Constipation	gt2_Convulsion	gt2_Cough	gt2_Crackles	gt2_Cramp	gt2_Crepitus	gt2_Croup	gt2_Cyanosis	gt2_Dandruff	gt2_Delayed onset muscle soreness	gt2_Dementia	gt2_Dentin hypersensitivity	gt2_Depersonalization	gt2_Depression	gt2_Dermatitis	gt2_Desquamation	gt2_Developmental disability	gt2_Diabetes	gt2_Diabetic ketoacidosis	gt2_Diarrhea	gt2_Dizziness	gt2_Dry eye syndrome	gt2_Dysautonomia	gt2_Dysgeusia	gt2_Dysmenorrhea	gt2_Dyspareunia	gt2_Dysphagia	gt2_Dysphoria	gt2_Dystonia	gt2_Dysuria	gt2_Ear pain	gt2_Eczema	gt2_Edema	gt2_Encephalitis	gt2_Encephalopathy	gt2_Epidermoid cyst	gt2_Epilepsy	gt2_Epiphora	gt2_Erectile dysfunction	gt2_Erythema	gt2_Erythema chronicum migrans	gt2_Esophagitis	gt2_Excessive daytime sleepiness	gt2_Eye pain	gt2_Eye strain	gt2_Facial nerve paralysis	gt2_Facial swelling	gt2_Fasciculation	gt2_Fatigue	gt2_Fatty liver disease	gt2_Fecal incontinence	gt2_Fever	gt2_Fibrillation	gt2_Fibrocystic breast changes	gt2_Fibromyalgia	gt2_Flatulence	gt2_Floater	gt2_Focal seizure	gt2_Folate deficiency	gt2_Food craving	gt2_Food intolerance	gt2_Frequent urination	gt2_Gastroesophageal reflux disease	gt2_Gastroparesis	gt2_Generalized anxiety disorder	gt2_Generalized tonic–clonic seizure	gt2_Genital wart	gt2_Gingival recession	gt2_Gingivitis	gt2_Globus pharyngis	gt2_Goitre	gt2_Gout	gt2_Grandiosity	gt2_Granuloma	gt2_Guilt	gt2_Hair loss	gt2_Halitosis	gt2_Hay fever	gt2_Headache	gt2_Heart arrhythmia	gt2_Heart murmur	gt2_Heartburn	gt2_Hematochezia	gt2_Hematoma	gt2_Hematuria	gt2_Hemolysis	gt2_Hemoptysis	gt2_Hemorrhoids	gt2_Hepatic encephalopathy	gt2_Hepatitis	gt2_Hepatotoxicity	gt2_Hiccup	gt2_Hip pain	gt2_Hives	gt2_Hot flash	gt2_Hydrocephalus	gt2_Hypercalcaemia	gt2_Hypercapnia	gt2_Hypercholesterolemia	gt2_Hyperemesis gravidarum	gt2_Hyperglycemia	gt2_Hyperkalemia	gt2_Hyperlipidemia	gt2_Hypermobility	gt2_Hyperpigmentation	gt2_Hypersomnia	gt2_Hypertension	gt2_Hyperthermia	gt2_Hyperthyroidism	gt2_Hypertriglyceridemia	gt2_Hypertrophy	gt2_Hyperventilation	gt2_Hypocalcaemia	gt2_Hypochondriasis	gt2_Hypoglycemia	gt2_Hypogonadism	gt2_Hypokalemia	gt2_Hypomania	gt2_Hyponatremia	gt2_Hypotension	gt2_Hypothyroidism	gt2_Hypoxemia	gt2_Hypoxia	gt2_Impetigo	gt2_Implantation bleeding	gt2_Impulsivity	gt2_Indigestion	gt2_Infection	gt2_Inflammation	gt2_Inflammatory bowel disease	gt2_Ingrown hair	gt2_Insomnia	gt2_Insulin resistance	gt2_Intermenstrual bleeding	gt2_Intracranial pressure	gt2_Iron deficiency	gt2_Irregular menstruation	gt2_Itch	gt2_Jaundice	gt2_Kidney failure	gt2_Kidney stone	gt2_Knee Pain	gt2_Kyphosis	gt2_Lactose intolerance	gt2_Laryngitis	gt2_Leg cramps	gt2_Lesion	gt2_Leukorrhea	gt2_Lightheadedness	gt2_Low back pain	gt2_Low-grade fever	gt2_Lymphedema	gt2_Major depressive disorder	gt2_Malabsorption	gt2_Male infertility	gt2_Manic Disorder	gt2_Melasma	gt2_Melena	gt2_Meningitis	gt2_Menorrhagia	gt2_Middle back pain	gt2_Migraine	gt2_Milium	gt2_Mitral insufficiency	gt2_Mood disorder	gt2_Mood swing	gt2_Morning sickness	gt2_Motion sickness	gt2_Mouth ulcer	gt2_Muscle atrophy	gt2_Muscle weakness	gt2_Myalgia	gt2_Mydriasis	gt2_Myocardial infarction	gt2_Myoclonus	gt2_Nasal congestion	gt2_Nasal polyp	gt2_Nausea	gt2_Neck mass	gt2_Neck pain	gt2_Neonatal jaundice	gt2_Nerve injury	gt2_Neuralgia	gt2_Neutropenia	gt2_Night sweats	gt2_Night terror	gt2_Nocturnal enuresis	gt2_Nodule	gt2_Nosebleed	gt2_Nystagmus	gt2_Obesity	gt2_Onychorrhexis	gt2_Oral candidiasis	gt2_Orthostatic hypotension	gt2_Osteopenia	gt2_Osteophyte	gt2_Osteoporosis	gt2_Otitis	gt2_Otitis externa	gt2_Otitis media	gt2_Pain	gt2_Palpitations	gt2_Panic attack	gt2_Papule	gt2_Paranoia	gt2_Paresthesia	gt2_Pelvic inflammatory disease	gt2_Pericarditis	gt2_Periodontal disease	gt2_Periorbital puffiness	gt2_Peripheral neuropathy	gt2_Perspiration	gt2_Petechia	gt2_Phlegm	gt2_Photodermatitis	gt2_Photophobia	gt2_Photopsia	gt2_Pleural effusion	gt2_Pleurisy	gt2_Pneumonia	gt2_Podalgia	gt2_Polycythemia	gt2_Polydipsia	gt2_Polyneuropathy	gt2_Polyuria	gt2_Poor posture	gt2_Post-nasal drip	gt2_Postural orthostatic tachycardia syndrome	gt2_Prediabetes	gt2_Proteinuria	gt2_Pruritus ani	gt2_Psychosis	gt2_Ptosis	gt2_Pulmonary edema	gt2_Pulmonary hypertension	gt2_Purpura	gt2_Pus	gt2_Pyelonephritis	gt2_Radiculopathy	gt2_Rectal pain	gt2_Rectal prolapse	gt2_Red eye	gt2_Renal colic	gt2_Restless legs syndrome	gt2_Rheum	gt2_Rhinitis	gt2_Rhinorrhea	gt2_Rosacea	gt2_Round ligament pain	gt2_Rumination	gt2_Scar	gt2_Sciatica	gt2_Scoliosis	gt2_Seborrheic dermatitis	gt2_Self-harm	gt2_Sensitivity to sound	gt2_Sexual dysfunction	gt2_Shallow breathing	gt2_Sharp pain	gt2_Shivering	gt2_Shortness of breath	gt2_Shyness	gt2_Sinusitis	gt2_Skin condition	gt2_Skin rash	gt2_Skin tag	gt2_Skin ulcer	gt2_Sleep apnea	gt2_Sleep deprivation	gt2_Sleep disorder	gt2_Snoring	gt2_Sore throat	gt2_Spasticity	gt2_Splenomegaly	gt2_Sputum	gt2_Stomach rumble	gt2_Strabismus	gt2_Stretch marks	gt2_Stridor	gt2_Stroke	gt2_Stuttering	gt2_Subdural hematoma	gt2_Suicidal ideation	gt2_Swelling	gt2_Swollen feet	gt2_Swollen lymph nodes	gt2_Syncope	gt2_Tachycardia	gt2_Tachypnea	gt2_Telangiectasia	gt2_Tenderness	gt2_Testicular pain	gt2_Throat irritation	gt2_Thrombocytopenia	gt2_Thyroid nodule	gt2_Tic	gt2_Tinnitus	gt2_Tonsillitis	gt2_Toothache	gt2_Tremor	gt2_Trichoptilosis	gt2_Tumor	gt2_Type 2 diabetes	gt2_Unconsciousness	gt2_Underweight	gt2_Upper respiratory tract infection	gt2_Urethritis	gt2_Urinary incontinence	gt2_Urinary tract infection	gt2_Urinary urgency	gt2_Uterine contraction	gt2_Vaginal bleeding	gt2_Vaginal discharge	gt2_Vaginitis	gt2_Varicose veins	gt2_Vasculitis	gt2_Ventricular fibrillation	gt2_Ventricular tachycardia	gt2_Vertigo	gt2_Viral pneumonia	gt2_Visual acuity	gt2_Vomiting	gt2_Wart	gt2_Water retention	gt2_Weakness	gt2_Weight gain	gt2_Wheeze	gt2_Xeroderma	gt2_Xerostomia	gt2_Yawn	gt2_hyperhidrosis	gt2_pancreatitis	neighbor_Alabama	neighbor_Alaska	neighbor_Arizona	neighbor_Arkansas	neighbor_California	neighbor_Colorado	neighbor_Connecticut	neighbor_Delaware	neighbor_Florida	neighbor_Georgia	neighbor_Hawaii	neighbor_Idaho	neighbor_Illinois	neighbor_Indiana	neighbor_Iowa	neighbor_Kansas	neighbor_Kentucky	neighbor_Louisiana	neighbor_Maine	neighbor_Maryland	neighbor_Massachusetts	neighbor_Michigan	neighbor_Minnesota	neighbor_Missouri	neighbor_Montana	neighbor_Nebraska	neighbor_Nevada	neighbor_New Hampshire	neighbor_New Jersey	neighbor_New Mexico	neighbor_New York	neighbor_North Carolina	neighbor_North Dakota	neighbor_Ohio	neighbor_Oklahoma	neighbor_Oregon	neighbor_Pennsylvania	neighbor_Rhode Island	neighbor_South Carolina	neighbor_South Dakota	neighbor_Tennessee	neighbor_Texas	neighbor_Utah	neighbor_Vermont	neighbor_Virginia	neighbor_Washington	neighbor_West Virginia	neighbor_Wisconsin	neighbor_Wyoming
0	2020-01-01	NaN	NaN	NaN	NaN	0.00000	0.000000	4486.684450	2.77	4.82	7.99	0.23	0.78	NaN	NaN	4.56	NaN	9.14	0.13	NaN	0.41	0.26	0.28	1.59	0.81	0.14	NaN	NaN	6.22	0.18	NaN	0.44	0.39	4.85	NaN	0.72	NaN	1.59	0.19	NaN	0.14	2.11	NaN	1.29	0.14	5.30	1.88	NaN	NaN	0.17	0.19	0.22	4.56	NaN	0.41	2.17	0.32	0.34	0.22	1.10	2.43	0.18	0.20	0.54	0.55	0.91	0.56	1.22	1.65	0.19	0.45	1.93	NaN	0.19	3.18	0.99	0.47	0.84	NaN	0.45	0.27	0.32	1.56	0.45	0.16	0.79	0.50	0.15	NaN	0.20	0.36	0.27	15.29	0.49	NaN	0.19	0.16	0.92	3.63	0.25	9.27	NaN	3.61	NaN	NaN	NaN	0.95	0.41	1.85	0.26	NaN	3.95	2.14	0.43	NaN	6.78	0.16	3.79	1.70	0.44	NaN	NaN	1.28	0.16	0.14	0.15	0.16	0.20	0.93	0.14	0.68	NaN	0.19	0.17	1.95	NaN	3.03	0.43	NaN	NaN	0.30	0.14	0.21	NaN	NaN	0.81	4.59	0.30	NaN	4.03	0.37	0.15	0.47	2.78	0.21	NaN	0.16	0.61	0.43	0.54	3.82	0.16	0.83	0.23	1.07	NaN	0.38	NaN	NaN	1.20	NaN	0.2	0.30	4.89	0.49	1.38	5.13	1.49	NaN	2.96	0.20	0.28	0.36	NaN	0.16	1.61	0.16	0.70	NaN	0.52	0.84	1.08	0.53	NaN	NaN	NaN	0.90	NaN	0.90	0.50	NaN	NaN	0.83	NaN	6.24	NaN	0.71	0.14	NaN	NaN	NaN	NaN	0.61	0.69	0.63	NaN	0.16	1.04	1.23	NaN	NaN	0.19	1.43	0.28	1.69	21.59	5.24	0.82	0.81	2.51	0.33	0.23	0.20	1.32	0.50	7.22	0.31	0.99	1.28	1.38	NaN	0.54	0.27	0.43	0.64	0.30	0.26	2.82	0.15	0.15	3.17	NaN	0.16	0.47	NaN	0.25	0.47	0.81	0.46	4.39	0.26	NaN	0.81	0.25	1.47	0.29	0.87	NaN	0.14	0.44	0.20	2.01	0.14	2.88	NaN	2.53	NaN	1.65	0.16	0.23	0.18	NaN	0.35	NaN	NaN	0.26	0.72	NaN	5.15	NaN	0.48	0.23	NaN	0.32	0.45	1.38	0.48	1.03	36.13	0.53	1.25	0.15	0.17	2.05	0.69	NaN	0.56	0.80	0.79	1.40	NaN	0.84	NaN	NaN	0.16	0.21	0.28	2.11	0.85	NaN	NaN	NaN	NaN	0.55	0.18	0.20	0.43	0.25	0.30	1.24	NaN	NaN	0.19	0.14	0.66	0.44	0.28	NaN	0.17	0.26	0.16	0.60	0.15	0.23	0.46	0.50	0.15	NaN	2.26	1.16	0.26	0.27	0.28	0.14	0.75	NaN	0.91	0.18	0.78	0.30	2.55	3.78	6.34	0.44	1.81	1.07	0.29	2.05	0.97	2.52	NaN	NaN	0.17	NaN	0.17	0.43	NaN	2.65	NaN	NaN	0.40	5.24	0.43	0.72	0.52	0.81	NaN	0.15	0.20	0.32	0.26	0.16	NaN	0.23	1.01	0.67	1.38	0.65	0.16	0.76	3.32	NaN	0.44	0.28	0.25	1.28	4.47	NaN	0.91	0.79	3.77	1.04	0.33	NaN	NaN	NaN	0.86	0.14	0.56	3.46	1.22	0.20	0.31	3.36	0.39	2.36	0.32	NaN	0.40	0.46	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	2020-01-02	NaN	NaN	NaN	NaN	1904.56187	1906.943228	1904.559254	2.71	4.98	8.64	0.26	0.85	NaN	NaN	3.74	NaN	10.44	NaN	0.20	0.31	0.26	0.32	1.93	0.99	0.22	NaN	NaN	6.84	0.19	NaN	0.41	0.31	5.08	NaN	0.76	0.18	2.37	0.26	NaN	0.18	2.92	NaN	1.38	NaN	5.83	2.16	NaN	NaN	0.38	0.27	0.18	4.46	0.13	0.52	2.26	0.49	0.36	0.16	1.11	2.65	0.18	0.38	0.57	0.39	0.86	0.65	1.29	1.41	0.25	0.41	1.72	NaN	NaN	3.57	0.91	0.46	0.93	0.12	0.54	0.31	0.39	1.65	0.39	NaN	0.99	0.64	0.17	0.12	0.22	0.52	0.27	15.25	0.58	0.19	0.18	0.22	0.97	3.56	0.26	9.74	NaN	3.43	NaN	0.20	NaN	1.05	0.46	1.85	0.19	NaN	4.20	2.32	0.36	0.18	8.24	0.23	3.82	1.66	0.51	0.12	0.17	1.48	0.23	0.35	0.14	0.12	0.24	0.84	0.27	0.92	0.22	0.24	0.31	2.43	NaN	3.52	0.39	NaN	0.12	0.28	0.17	0.23	0.26	0.14	0.82	4.88	0.46	0.13	4.05	0.48	0.13	0.62	2.41	0.20	0.17	0.22	0.60	0.58	0.75	4.00	0.18	1.03	0.13	1.10	0.12	0.42	NaN	0.13	1.19	0.13	NaN	0.28	4.76	0.60	1.59	4.68	1.95	0.18	3.29	0.21	0.29	0.53	NaN	0.15	1.54	0.16	0.95	0.17	0.46	0.93	1.06	0.83	0.15	NaN	NaN	1.47	NaN	1.04	0.64	0.19	0.17	0.80	NaN	8.67	0.23	1.03	0.25	0.24	NaN	NaN	NaN	0.64	0.76	0.64	NaN	0.26	1.25	1.64	0.19	0.13	0.23	1.49	0.41	1.63	25.42	5.73	0.94	0.81	2.95	0.35	0.29	0.23	1.42	0.55	7.43	0.46	1.32	1.46	1.40	NaN	0.66	0.29	0.37	0.84	0.20	0.52	2.91	0.16	0.24	3.99	0.15	0.15	0.52	0.17	0.32	0.71	0.93	0.48	4.72	0.19	NaN	0.99	0.17	1.28	0.32	0.74	NaN	0.27	0.58	0.23	2.38	0.14	2.70	NaN	2.44	0.13	1.55	0.20	0.28	0.20	NaN	0.35	NaN	0.22	0.29	0.70	NaN	6.06	NaN	0.65	0.29	0.24	0.21	0.76	1.50	0.56	1.08	37.69	0.47	1.32	0.12	0.14	1.99	0.83	0.14	0.75	0.66	0.86	1.68	NaN	0.80	NaN	NaN	0.13	0.33	0.19	2.07	0.99	0.20	NaN	0.14	0.13	0.60	0.24	0.20	0.56	0.23	0.34	1.59	0.14	0.13	0.37	0.17	0.71	0.59	0.37	NaN	0.17	0.31	0.16	0.63	0.12	0.23	0.51	0.37	0.19	NaN	2.79	1.17	0.30	0.29	0.21	NaN	0.77	NaN	0.88	0.16	0.82	0.21	2.67	4.38	6.49	0.41	1.87	1.32	0.36	2.42	0.91	2.69	NaN	0.12	0.22	NaN	0.26	0.45	NaN	2.66	0.14	NaN	0.36	5.24	0.53	1.05	0.65	0.86	NaN	0.17	0.19	0.32	0.33	0.37	0.14	0.20	0.90	0.73	1.18	0.60	0.22	1.07	4.47	0.22	0.58	0.32	0.35	1.41	5.18	NaN	0.84	0.75	3.93	1.15	0.33	NaN	NaN	0.17	1.07	0.19	0.45	3.10	1.36	0.21	0.49	3.56	0.42	2.41	0.37	0.16	0.45	0.33	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	2020-01-03	NaN	NaN	NaN	NaN	0.00000	0.000000	3842.010974	2.44	4.85	7.91	0.35	0.80	0.12	NaN	3.50	NaN	9.60	0.18	0.17	0.27	0.26	0.39	2.00	0.89	0.23	NaN	NaN	6.74	0.15	0.12	0.46	0.34	4.60	0.21	0.84	0.12	2.12	0.25	0.19	0.13	2.84	NaN	1.25	NaN	5.34	2.07	0.12	NaN	0.18	0.25	0.17	4.17	0.14	0.49	2.15	0.45	0.42	0.16	1.06	2.79	0.17	0.17	0.41	0.53	0.85	0.61	1.13	1.47	0.19	0.43	1.67	NaN	NaN	3.53	0.91	0.32	0.73	NaN	0.54	0.23	0.31	1.47	0.35	NaN	0.85	0.66	NaN	NaN	0.23	0.44	0.28	14.37	0.53	0.14	0.21	0.20	1.00	3.40	0.32	9.18	0.12	3.40	NaN	0.19	0.15	1.19	0.45	1.74	0.25	NaN	3.71	2.27	0.36	NaN	8.24	0.21	3.67	1.51	0.59	NaN	NaN	1.49	NaN	0.28	0.19	NaN	0.23	0.85	0.16	0.83	0.20	0.17	0.28	2.31	NaN	3.44	0.45	NaN	NaN	0.22	0.22	0.23	0.18	0.19	0.74	4.35	0.46	0.19	3.87	0.41	0.15	0.48	2.92	0.20	NaN	NaN	0.57	0.50	0.59	4.07	NaN	1.01	0.24	1.10	0.14	0.31	NaN	NaN	1.05	0.14	NaN	0.15	4.48	0.68	1.57	4.12	1.85	NaN	3.10	0.26	0.37	0.39	NaN	0.17	1.52	NaN	0.87	0.12	0.47	0.84	1.13	0.73	0.12	NaN	NaN	1.42	0.14	1.14	0.56	0.20	0.12	0.79	NaN	7.81	0.16	0.90	0.23	0.16	NaN	0.13	NaN	0.68	0.69	0.72	NaN	0.25	1.20	1.54	0.17	0.17	0.19	1.54	0.29	1.69	24.01	5.54	0.90	0.75	2.97	0.31	0.28	0.23	1.35	0.53	7.07	0.42	1.29	1.48	1.27	NaN	0.62	0.12	0.42	0.95	0.24	0.46	2.89	0.16	0.22	3.70	0.24	0.18	0.61	0.15	0.26	0.56	0.69	0.43	4.26	0.24	0.13	0.93	0.21	1.36	0.36	0.63	NaN	0.30	0.51	NaN	2.27	NaN	2.63	NaN	2.34	0.18	1.47	NaN	0.27	0.20	0.13	0.40	NaN	NaN	0.38	0.57	NaN	5.43	0.20	0.53	0.31	0.14	0.25	0.73	1.41	0.46	0.99	35.84	0.54	1.30	NaN	0.13	1.84	0.82	NaN	0.58	0.72	0.80	1.81	NaN	0.83	NaN	NaN	0.15	0.21	0.23	2.12	0.77	0.15	NaN	0.18	0.12	0.55	0.24	0.22	0.56	0.15	0.30	1.59	NaN	NaN	0.31	0.25	0.68	0.60	0.37	NaN	0.12	0.34	0.21	0.60	0.14	0.22	0.45	0.36	0.16	0.15	2.41	1.18	0.40	0.36	0.20	NaN	0.66	NaN	0.95	0.16	0.72	0.21	2.45	4.14	6.04	0.47	1.78	1.28	0.23	2.30	0.90	2.42	NaN	0.15	0.25	0.13	0.31	0.47	NaN	2.50	0.17	NaN	0.29	5.12	0.53	1.05	0.47	0.85	0.12	0.17	0.21	0.33	0.24	0.36	0.13	0.24	0.90	0.74	1.31	0.58	0.12	0.86	4.48	0.15	0.54	0.34	0.32	1.39	5.21	0.16	0.94	0.69	3.63	1.06	0.36	NaN	NaN	0.23	0.89	0.27	0.35	3.04	1.36	0.31	0.43	3.06	0.35	2.01	0.29	0.15	0.41	0.24	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	2020-01-04	NaN	NaN	NaN	NaN	0.00000	0.000000	7949.995496	2.71	4.80	8.31	0.22	0.67	NaN	NaN	3.51	NaN	9.22	NaN	NaN	0.34	0.28	0.26	1.62	0.84	0.20	0.17	NaN	6.34	0.19	NaN	0.39	0.32	4.81	NaN	0.85	NaN	1.72	0.21	NaN	NaN	2.30	NaN	1.23	NaN	4.90	2.01	0.13	NaN	0.20	0.23	0.17	3.82	0.15	0.46	2.18	0.48	0.30	0.18	1.13	2.33	NaN	0.19	0.46	0.56	0.85	0.52	1.06	1.31	0.26	0.43	1.64	NaN	0.15	3.36	1.11	0.38	0.82	NaN	0.43	0.22	0.32	1.55	0.49	NaN	0.72	0.67	NaN	NaN	0.22	0.49	0.20	14.13	0.44	NaN	0.24	NaN	0.95	3.80	0.21	8.22	NaN	3.71	NaN	0.19	NaN	1.31	0.40	1.81	0.20	NaN	3.59	2.46	0.45	NaN	7.01	0.21	3.85	1.47	0.60	NaN	NaN	1.46	0.13	0.33	NaN	NaN	0.14	0.90	0.16	0.65	0.18	0.15	0.22	1.91	NaN	3.29	0.44	NaN	NaN	0.29	0.21	0.21	0.14	0.13	0.77	4.39	0.42	0.14	3.66	0.38	NaN	0.57	2.53	0.17	NaN	0.20	0.51	0.40	0.55	3.99	0.24	0.83	0.13	1.02	NaN	0.36	NaN	NaN	1.11	NaN	NaN	0.39	5.18	0.62	1.42	4.18	1.61	NaN	3.21	0.25	0.23	0.36	NaN	NaN	1.59	NaN	0.65	NaN	0.68	0.76	1.01	0.69	0.17	NaN	NaN	1.02	NaN	0.92	0.50	NaN	0.15	0.81	NaN	6.52	NaN	0.77	NaN	0.15	NaN	NaN	NaN	0.60	0.76	0.63	NaN	0.16	1.08	1.37	0.13	0.14	0.21	1.31	0.26	1.53	22.57	5.27	0.55	0.85	2.86	0.30	0.23	0.22	1.25	0.43	7.38	0.35	1.05	1.38	1.28	NaN	0.61	NaN	0.56	0.76	0.28	0.57	2.72	0.14	0.18	3.12	NaN	NaN	0.51	0.17	0.26	0.60	0.69	0.38	4.07	0.23	NaN	0.67	0.20	1.11	0.18	0.80	0.17	0.22	0.46	0.19	1.92	NaN	2.60	NaN	2.46	NaN	1.50	0.16	0.22	0.22	NaN	0.34	NaN	0.16	0.25	0.49	NaN	4.74	NaN	0.50	0.22	NaN	0.25	0.53	1.48	0.46	1.01	35.48	0.50	1.37	0.15	0.14	1.81	0.62	NaN	0.54	0.80	0.74	1.72	NaN	0.81	NaN	0.15	0.14	0.21	0.23	1.81	0.93	0.18	NaN	0.18	0.13	0.61	0.18	0.15	0.46	0.23	0.30	1.44	NaN	NaN	0.23	0.18	0.66	0.55	0.32	0.13	0.22	0.27	0.14	0.67	0.14	0.25	0.43	0.30	NaN	NaN	2.48	1.05	0.29	0.40	0.21	NaN	0.72	NaN	0.97	0.18	0.69	0.25	2.38	3.84	6.17	0.45	1.64	1.00	0.38	2.49	0.85	2.16	NaN	0.16	0.23	0.14	0.25	0.62	NaN	2.08	0.16	NaN	0.35	4.96	0.50	0.85	0.49	0.70	NaN	0.17	0.21	0.29	0.33	0.29	NaN	0.21	0.92	0.73	1.19	0.63	0.20	0.79	3.68	0.23	0.53	0.30	0.28	1.12	4.67	NaN	1.00	0.78	3.71	1.05	0.33	NaN	NaN	0.13	0.99	NaN	0.28	3.03	1.19	0.22	0.43	3.12	0.33	2.53	0.37	0.23	0.44	0.29	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	2020-01-05	NaN	NaN	NaN	NaN	0.00000	0.000000	6054.562803	2.90	4.99	8.58	0.23	0.65	NaN	NaN	3.30	NaN	8.81	0.16	NaN	0.38	0.29	0.38	1.57	0.88	NaN	NaN	NaN	6.72	NaN	NaN	0.36	0.33	4.77	0.16	0.80	NaN	1.55	0.24	NaN	NaN	2.48	NaN	1.21	NaN	5.11	1.80	NaN	0.17	0.33	0.18	0.17	3.99	0.23	0.62	2.40	0.41	0.29	0.25	1.24	3.42	0.34	0.38	0.46	0.64	0.93	0.69	0.93	1.60	NaN	0.35	1.79	NaN	0.25	3.49	0.97	0.46	0.68	NaN	0.40	0.22	0.28	1.55	0.37	NaN	0.73	0.55	NaN	NaN	0.25	0.41	0.21	13.39	0.42	NaN	0.24	NaN	0.97	3.83	0.24	8.17	NaN	3.71	NaN	0.18	NaN	1.40	0.40	1.77	0.29	NaN	4.02	2.45	0.41	NaN	7.33	0.21	3.50	1.63	0.47	NaN	NaN	1.41	0.13	0.23	NaN	NaN	0.27	0.72	NaN	0.75	NaN	0.23	0.20	2.10	NaN	2.78	0.45	NaN	NaN	0.30	0.14	0.24	0.20	0.13	0.80	4.74	0.42	0.14	3.71	0.41	NaN	0.56	2.56	0.19	NaN	0.16	0.65	0.55	0.63	3.83	0.31	1.04	0.16	1.16	0.14	0.38	0.18	NaN	1.01	NaN	NaN	0.31	4.90	0.61	1.46	4.65	1.56	0.15	3.04	0.20	0.23	0.39	NaN	0.18	1.41	NaN	0.60	NaN	0.52	0.74	1.08	0.66	NaN	NaN	NaN	0.99	NaN	1.00	0.43	0.15	0.14	0.95	NaN	6.84	NaN	0.76	0.15	0.19	NaN	NaN	NaN	0.62	0.75	0.54	NaN	0.18	1.13	1.41	0.17	NaN	0.18	1.53	0.23	1.49	22.45	5.30	0.79	0.85	3.06	0.34	0.40	0.23	1.27	0.50	7.56	0.34	1.04	1.32	1.40	NaN	0.69	0.13	0.42	0.72	0.29	0.39	2.72	0.17	0.18	3.05	NaN	NaN	0.48	0.15	0.16	0.42	0.86	0.31	4.25	0.24	NaN	0.94	0.20	1.21	0.24	0.82	NaN	NaN	0.50	0.13	2.04	NaN	2.44	NaN	2.35	NaN	1.55	0.13	0.28	0.24	NaN	0.35	NaN	NaN	0.18	0.49	0.13	4.91	0.17	0.51	0.29	NaN	0.21	0.52	1.39	0.34	0.82	36.06	0.42	1.36	0.13	0.17	1.90	0.65	NaN	0.71	0.83	0.68	1.82	NaN	0.80	NaN	NaN	NaN	0.18	0.35	1.62	0.92	NaN	NaN	0.14	0.18	0.54	0.17	NaN	0.60	0.20	0.18	1.44	NaN	NaN	0.23	NaN	0.71	0.61	0.34	NaN	0.14	0.25	0.14	0.60	0.13	0.20	0.51	0.43	0.18	NaN	2.86	1.07	0.25	0.25	0.22	NaN	0.66	NaN	0.98	0.19	0.69	0.28	2.25	3.84	6.10	0.44	1.84	0.98	0.33	2.52	0.99	2.34	NaN	NaN	0.19	0.14	0.26	0.53	0.14	2.15	0.15	0.16	0.31	5.25	0.47	0.92	0.53	0.80	0.13	0.23	0.21	0.36	0.28	0.18	NaN	0.18	0.91	0.68	1.29	0.60	0.16	0.90	3.93	0.23	0.55	0.16	0.23	1.25	4.88	NaN	0.85	0.74	3.75	1.08	0.36	NaN	0.18	0.17	0.98	0.19	0.29	3.06	1.46	0.27	0.39	3.45	0.25	2.38	0.35	NaN	0.57	0.28	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

	date	up2date	gt_chest pain	gt_nose bleed	gt_fever	gt2_Abdominal obesity	gt2_Abdominal pain	gt2_Acne	gt2_Actinic keratosis	gt2_Acute bronchitis	gt2_Adrenal crisis	gt2_Ageusia	gt2_Alcoholism	gt2_Allergic conjunctivitis	gt2_Allergy	gt2_Amblyopia	gt2_Amenorrhea	gt2_Amnesia	gt2_Anal fissure	gt2_Anaphylaxis	gt2_Anemia	gt2_Angina pectoris	gt2_Angioedema	gt2_Angular cheilitis	gt2_Anosmia	gt2_Anxiety	gt2_Aphasia	gt2_Aphonia	gt2_Apnea	gt2_Arthralgia	gt2_Arthritis	gt2_Ascites	gt2_Asperger syndrome	gt2_Asphyxia	gt2_Asthma	gt2_Astigmatism	gt2_Ataxia	gt2_Atheroma	gt2_Attention deficit hyperactivity disorder	gt2_Auditory hallucination	gt2_Autoimmune disease	gt2_Avoidant personality disorder	gt2_Back pain	gt2_Bacterial vaginosis	gt2_Balance disorder	gt2_Beau's lines	gt2_Bell's palsy	gt2_Biliary colic	gt2_Binge eating	gt2_Bleeding	gt2_Bleeding on probing	gt2_Blepharospasm	gt2_Bloating	gt2_Blood in stool	gt2_Blurred vision	gt2_Blushing	gt2_Boil	gt2_Bone fracture	gt2_Bone tumor	gt2_Bowel obstruction	gt2_Bradycardia	gt2_Braxton Hicks contractions	gt2_Breakthrough bleeding	gt2_Breast pain	gt2_Bronchitis	gt2_Bruise	gt2_Bruxism	gt2_Bunion	gt2_Burn	gt2_Burning Chest Pain	gt2_Burning mouth syndrome	gt2_Candidiasis	gt2_Canker sore	gt2_Cardiac arrest	gt2_Carpal tunnel syndrome	gt2_Cataplexy	gt2_Cataract	gt2_Chancre	gt2_Cheilitis	gt2_Chest pain	gt2_Chills	gt2_Chorea	gt2_Chronic pain	gt2_Cirrhosis	gt2_Cleft lip and cleft palate	gt2_Clouding of consciousness	gt2_Cluster headache	gt2_Colitis	gt2_Coma	gt2_Common cold	gt2_Compulsive behavior	gt2_Compulsive hoarding	gt2_Confusion	gt2_Congenital heart defect	gt2_Conjunctivitis	gt2_Constipation	gt2_Convulsion	gt2_Cough	gt2_Crackles	gt2_Cramp	gt2_Crepitus	gt2_Croup	gt2_Cyanosis	gt2_Dandruff	gt2_Delayed onset muscle soreness	gt2_Dementia	gt2_Dentin hypersensitivity	gt2_Depersonalization	gt2_Depression	gt2_Dermatitis	gt2_Desquamation	gt2_Developmental disability	gt2_Diabetes	gt2_Diabetic ketoacidosis	gt2_Diarrhea	gt2_Dizziness	gt2_Dry eye syndrome	gt2_Dysautonomia	gt2_Dysgeusia	gt2_Dysmenorrhea	gt2_Dyspareunia	gt2_Dysphagia	gt2_Dysphoria	gt2_Dystonia	gt2_Dysuria	gt2_Ear pain	gt2_Eczema	gt2_Edema	gt2_Encephalitis	gt2_Encephalopathy	gt2_Epidermoid cyst	gt2_Epilepsy	gt2_Epiphora	gt2_Erectile dysfunction	gt2_Erythema	gt2_Erythema chronicum migrans	gt2_Esophagitis	gt2_Excessive daytime sleepiness	gt2_Eye pain	gt2_Eye strain	gt2_Facial nerve paralysis	gt2_Facial swelling	gt2_Fasciculation	gt2_Fatigue	gt2_Fatty liver disease	gt2_Fecal incontinence	gt2_Fever	gt2_Fibrillation	gt2_Fibrocystic breast changes	gt2_Fibromyalgia	gt2_Flatulence	gt2_Floater	gt2_Focal seizure	gt2_Folate deficiency	gt2_Food craving	gt2_Food intolerance	gt2_Frequent urination	gt2_Gastroesophageal reflux disease	gt2_Gastroparesis	gt2_Generalized anxiety disorder	gt2_Generalized tonic–clonic seizure	gt2_Genital wart	gt2_Gingival recession	gt2_Gingivitis	gt2_Globus pharyngis	gt2_Goitre	gt2_Gout	gt2_Grandiosity	gt2_Granuloma	gt2_Guilt	gt2_Hair loss	gt2_Halitosis	gt2_Hay fever	gt2_Headache	gt2_Heart arrhythmia	gt2_Heart murmur	gt2_Heartburn	gt2_Hematochezia	gt2_Hematoma	gt2_Hematuria	gt2_Hemolysis	gt2_Hemoptysis	gt2_Hemorrhoids	gt2_Hepatic encephalopathy	gt2_Hepatitis	gt2_Hepatotoxicity	gt2_Hiccup	gt2_Hip pain	gt2_Hives	gt2_Hot flash	gt2_Hydrocephalus	gt2_Hypercalcaemia	gt2_Hypercapnia	gt2_Hypercholesterolemia	gt2_Hyperemesis gravidarum	gt2_Hyperglycemia	gt2_Hyperkalemia	gt2_Hyperlipidemia	gt2_Hypermobility	gt2_Hyperpigmentation	gt2_Hypersomnia	gt2_Hypertension	gt2_Hyperthermia	gt2_Hyperthyroidism	gt2_Hypertriglyceridemia	gt2_Hypertrophy	gt2_Hyperventilation	gt2_Hypocalcaemia	gt2_Hypochondriasis	gt2_Hypoglycemia	gt2_Hypogonadism	gt2_Hypokalemia	gt2_Hypomania	gt2_Hyponatremia	gt2_Hypotension	gt2_Hypothyroidism	gt2_Hypoxemia	gt2_Hypoxia	gt2_Impetigo	gt2_Implantation bleeding	gt2_Impulsivity	gt2_Indigestion	gt2_Infection	gt2_Inflammation	gt2_Inflammatory bowel disease	gt2_Ingrown hair	gt2_Insomnia	gt2_Insulin resistance	gt2_Intermenstrual bleeding	gt2_Intracranial pressure	gt2_Iron deficiency	gt2_Irregular menstruation	gt2_Itch	gt2_Jaundice	gt2_Kidney failure	gt2_Kidney stone	gt2_Knee Pain	gt2_Kyphosis	gt2_Lactose intolerance	gt2_Laryngitis	gt2_Leg cramps	gt2_Lesion	gt2_Leukorrhea	gt2_Lightheadedness	gt2_Low back pain	gt2_Low-grade fever	gt2_Lymphedema	gt2_Major depressive disorder	gt2_Malabsorption	gt2_Male infertility	gt2_Manic Disorder	gt2_Melasma	gt2_Melena	gt2_Meningitis	gt2_Menorrhagia	gt2_Middle back pain	gt2_Migraine	gt2_Milium	gt2_Mitral insufficiency	gt2_Mood disorder	gt2_Mood swing	gt2_Morning sickness	gt2_Motion sickness	gt2_Mouth ulcer	gt2_Muscle atrophy	gt2_Muscle weakness	gt2_Myalgia	gt2_Mydriasis	gt2_Myocardial infarction	gt2_Myoclonus	gt2_Nasal congestion	gt2_Nasal polyp	gt2_Nausea	gt2_Neck mass	gt2_Neck pain	gt2_Neonatal jaundice	gt2_Nerve injury	gt2_Neuralgia	gt2_Neutropenia	gt2_Night sweats	gt2_Night terror	gt2_Nocturnal enuresis	gt2_Nodule	gt2_Nosebleed	gt2_Nystagmus	gt2_Obesity	gt2_Onychorrhexis	gt2_Oral candidiasis	gt2_Orthostatic hypotension	gt2_Osteopenia	gt2_Osteophyte	gt2_Osteoporosis	gt2_Otitis	gt2_Otitis externa	gt2_Otitis media	gt2_Pain	gt2_Palpitations	gt2_Panic attack	gt2_Papule	gt2_Paranoia	gt2_Paresthesia	gt2_Pelvic inflammatory disease	gt2_Pericarditis	gt2_Periodontal disease	gt2_Periorbital puffiness	gt2_Peripheral neuropathy	gt2_Perspiration	gt2_Petechia	gt2_Phlegm	gt2_Photodermatitis	gt2_Photophobia	gt2_Photopsia	gt2_Pleural effusion	gt2_Pleurisy	gt2_Pneumonia	gt2_Podalgia	gt2_Polycythemia	gt2_Polydipsia	gt2_Polyneuropathy	gt2_Polyuria	gt2_Poor posture	gt2_Post-nasal drip	gt2_Postural orthostatic tachycardia syndrome	gt2_Prediabetes	gt2_Proteinuria	gt2_Pruritus ani	gt2_Psychosis	gt2_Ptosis	gt2_Pulmonary edema	gt2_Pulmonary hypertension	gt2_Purpura	gt2_Pus	gt2_Pyelonephritis	gt2_Radiculopathy	gt2_Rectal pain	gt2_Rectal prolapse	gt2_Red eye	gt2_Renal colic	gt2_Restless legs syndrome	gt2_Rheum	gt2_Rhinitis	gt2_Rhinorrhea	gt2_Rosacea	gt2_Round ligament pain	gt2_Rumination	gt2_Scar	gt2_Sciatica	gt2_Scoliosis	gt2_Seborrheic dermatitis	gt2_Self-harm	gt2_Sensitivity to sound	gt2_Sexual dysfunction	gt2_Shallow breathing	gt2_Sharp pain	gt2_Shivering	gt2_Shortness of breath	gt2_Shyness	gt2_Sinusitis	gt2_Skin condition	gt2_Skin rash	gt2_Skin tag	gt2_Skin ulcer	gt2_Sleep apnea	gt2_Sleep deprivation	gt2_Sleep disorder	gt2_Snoring	gt2_Sore throat	gt2_Spasticity	gt2_Splenomegaly	gt2_Sputum	gt2_Stomach rumble	gt2_Strabismus	gt2_Stretch marks	gt2_Stridor	gt2_Stroke	gt2_Stuttering	gt2_Subdural hematoma	gt2_Suicidal ideation	gt2_Swelling	gt2_Swollen feet	gt2_Swollen lymph nodes	gt2_Syncope	gt2_Tachycardia	gt2_Tachypnea	gt2_Telangiectasia	gt2_Tenderness	gt2_Testicular pain	gt2_Throat irritation	gt2_Thrombocytopenia	gt2_Thyroid nodule	gt2_Tic	gt2_Tinnitus	gt2_Tonsillitis	gt2_Toothache	gt2_Tremor	gt2_Trichoptilosis	gt2_Tumor	gt2_Type 2 diabetes	gt2_Unconsciousness	gt2_Underweight	gt2_Upper respiratory tract infection	gt2_Urethritis	gt2_Urinary incontinence	gt2_Urinary tract infection	gt2_Urinary urgency	gt2_Uterine contraction	gt2_Vaginal bleeding	gt2_Vaginal discharge	gt2_Vaginitis	gt2_Varicose veins	gt2_Vasculitis	gt2_Ventricular fibrillation	gt2_Ventricular tachycardia	gt2_Vertigo	gt2_Viral pneumonia	gt2_Visual acuity	gt2_Vomiting	gt2_Wart	gt2_Water retention	gt2_Weakness	gt2_Weight gain	gt2_Wheeze	gt2_Xeroderma	gt2_Xerostomia	gt2_Yawn	gt2_hyperhidrosis	gt2_pancreatitis	neighbor_Washington	State_Name_Mississippi
0	2020-01-01	11.380448	0.00000	0.000000	4486.684450	2.77	4.82	7.99	0.23	0.78	0.120000	0.12	4.56	0.11	9.14	0.130	0.200000	0.41	0.26	0.28	1.59	0.81	0.14	0.17	0.11	6.22	0.180	0.1200	0.44	0.39	4.85	0.210	0.72	0.18	1.59	0.19	0.19	0.1400	2.11	0.15	1.29	0.140	5.30	1.88	0.12	0.17	0.17	0.19	0.22	4.56	0.13	0.41	2.17	0.32	0.34	0.22	1.10	2.43	0.180	0.20	0.54	0.55	0.91	0.56	1.22	1.65	0.190	0.45	1.93	0.11	0.190000	3.18	0.99	0.47	0.84	0.120000	0.45	0.27	0.32	1.56	0.45	0.160	0.79	0.50	0.15	0.120	0.20	0.36	0.27	15.29	0.49	0.190	0.19	0.160000	0.92	3.63	0.25	9.27	0.120	3.61	0.11	0.20	0.150000	0.95	0.41	1.85	0.26	0.19	3.95	2.14	0.43	0.18	6.78	0.16	3.79	1.70	0.44	0.12	0.170000	1.28	0.16	0.14	0.15	0.160	0.20	0.93	0.14	0.68	0.22	0.19	0.17	1.95	0.11	3.03	0.43	0.15	0.12	0.30	0.14	0.21	0.26	0.14	0.81	4.59	0.30	0.13	4.03	0.37	0.150000	0.47	2.78	0.21	0.17	0.16	0.61	0.43	0.54	3.82	0.16	0.83	0.23	1.07	0.12	0.38	0.18	0.130	1.20	0.130	0.200	0.30	4.89	0.49	1.38	5.13	1.49	0.18	2.96	0.20	0.28	0.36	0.11	0.160	1.61	0.160	0.70	0.1700	0.52	0.84	1.08	0.53	0.150	0.13	0.11	0.90	0.140000	0.90	0.50	0.190	0.17	0.83	0.25	6.24	0.230000	0.71	0.14	0.24	0.13	0.130	0.12	0.61	0.69	0.63	0.18	0.16	1.04	1.23	0.19	0.130	0.19	1.43	0.28	1.69	21.59	5.24	0.82	0.81	2.51	0.33	0.23	0.20	1.32	0.50	7.22	0.31	0.99	1.28	1.38	0.11	0.54	0.270	0.43	0.64	0.30	0.26	2.82	0.15	0.15	3.17	0.15	0.16	0.47	0.17	0.25	0.47	0.81	0.46	4.39	0.26	0.13	0.81	0.25	1.47	0.29	0.87	0.17	0.14	0.44	0.20	2.01	0.140	2.88	0.13	2.53	0.13	1.65	0.16	0.23	0.18	0.13	0.35	0.11	0.2200	0.26	0.72	0.13	5.15	0.200	0.48	0.23	0.240000	0.32	0.45	1.38	0.48	1.03	36.13	0.53	1.25	0.150	0.17	2.05	0.69	0.140	0.56	0.80	0.79	1.40	0.11	0.84	0.14	0.15	0.160	0.21	0.28	2.11	0.85	0.20	0.11	0.14	0.13	0.55	0.18	0.20	0.43	0.25	0.30	1.24	0.140000	0.130	0.19	0.14	0.66	0.44	0.28	0.13	0.17	0.26	0.16	0.60	0.15	0.23	0.46	0.50	0.15	0.15	2.26	1.16	0.26	0.27	0.28	0.140	0.75	0.11	0.91	0.18	0.78	0.30	2.55	3.78	6.34	0.44	1.81	1.07	0.29	2.05	0.97	2.52	0.15	0.12	0.17	0.13	0.17	0.43	0.14	2.65	0.14	0.16	0.40	5.24	0.43	0.72	0.52	0.81	0.120	0.15	0.20	0.32	0.26	0.16	0.14	0.23	1.01	0.67	1.38	0.65	0.16	0.76	3.32	0.22	0.44	0.28	0.25	1.28	4.47	0.160000	0.91	0.79	3.77	1.04	0.33	0.18	0.18	0.17	0.86	0.14	0.56	3.46	1.22	0.20	0.31	3.36	0.39	2.36	0.32	0.160	0.40	0.46	1.0	1
1	2020-01-02	11.380448	1904.56187	1906.943228	1904.559254	2.71	4.98	8.64	0.26	0.85	0.120000	0.12	3.74	0.11	10.44	0.155	0.200000	0.31	0.26	0.32	1.93	0.99	0.22	0.17	0.11	6.84	0.190	0.1200	0.41	0.31	5.08	0.210	0.76	0.18	2.37	0.26	0.19	0.1800	2.92	0.15	1.38	0.136	5.83	2.16	0.12	0.17	0.38	0.27	0.18	4.46	0.13	0.52	2.26	0.49	0.36	0.16	1.11	2.65	0.180	0.38	0.57	0.39	0.86	0.65	1.29	1.41	0.250	0.41	1.72	0.11	0.176667	3.57	0.91	0.46	0.93	0.120000	0.54	0.31	0.39	1.65	0.39	0.152	0.99	0.64	0.17	0.120	0.22	0.52	0.27	15.25	0.58	0.190	0.18	0.220000	0.97	3.56	0.26	9.74	0.120	3.43	0.11	0.20	0.150000	1.05	0.46	1.85	0.19	0.19	4.20	2.32	0.36	0.18	8.24	0.23	3.82	1.66	0.51	0.12	0.170000	1.48	0.23	0.35	0.14	0.120	0.24	0.84	0.27	0.92	0.22	0.24	0.31	2.43	0.11	3.52	0.39	0.15	0.12	0.28	0.17	0.23	0.26	0.14	0.82	4.88	0.46	0.13	4.05	0.48	0.130000	0.62	2.41	0.20	0.17	0.22	0.60	0.58	0.75	4.00	0.18	1.03	0.13	1.10	0.12	0.42	0.18	0.130	1.19	0.130	0.188	0.28	4.76	0.60	1.59	4.68	1.95	0.18	3.29	0.21	0.29	0.53	0.11	0.150	1.54	0.160	0.95	0.1700	0.46	0.93	1.06	0.83	0.150	0.13	0.11	1.47	0.140000	1.04	0.64	0.190	0.17	0.80	0.25	8.67	0.230000	1.03	0.25	0.24	0.13	0.130	0.12	0.64	0.76	0.64	0.18	0.26	1.25	1.64	0.19	0.130	0.23	1.49	0.41	1.63	25.42	5.73	0.94	0.81	2.95	0.35	0.29	0.23	1.42	0.55	7.43	0.46	1.32	1.46	1.40	0.11	0.66	0.290	0.37	0.84	0.20	0.52	2.91	0.16	0.24	3.99	0.15	0.15	0.52	0.17	0.32	0.71	0.93	0.48	4.72	0.19	0.13	0.99	0.17	1.28	0.32	0.74	0.17	0.27	0.58	0.23	2.38	0.140	2.70	0.13	2.44	0.13	1.55	0.20	0.28	0.20	0.13	0.35	0.11	0.2200	0.29	0.70	0.13	6.06	0.200	0.65	0.29	0.240000	0.21	0.76	1.50	0.56	1.08	37.69	0.47	1.32	0.120	0.14	1.99	0.83	0.140	0.75	0.66	0.86	1.68	0.11	0.80	0.14	0.15	0.130	0.33	0.19	2.07	0.99	0.20	0.11	0.14	0.13	0.60	0.24	0.20	0.56	0.23	0.34	1.59	0.140000	0.130	0.37	0.17	0.71	0.59	0.37	0.13	0.17	0.31	0.16	0.63	0.12	0.23	0.51	0.37	0.19	0.15	2.79	1.17	0.30	0.29	0.21	0.138	0.77	0.11	0.88	0.16	0.82	0.21	2.67	4.38	6.49	0.41	1.87	1.32	0.36	2.42	0.91	2.69	0.15	0.12	0.22	0.13	0.26	0.45	0.14	2.66	0.14	0.16	0.36	5.24	0.53	1.05	0.65	0.86	0.120	0.17	0.19	0.32	0.33	0.37	0.14	0.20	0.90	0.73	1.18	0.60	0.22	1.07	4.47	0.22	0.58	0.32	0.35	1.41	5.18	0.160000	0.84	0.75	3.93	1.15	0.33	0.18	0.18	0.17	1.07	0.19	0.45	3.10	1.36	0.21	0.49	3.56	0.42	2.41	0.37	0.160	0.45	0.33	1.0	1
2	2020-01-03	11.380448	0.00000	0.000000	3842.010974	2.44	4.85	7.91	0.35	0.80	0.120000	0.12	3.50	0.11	9.60	0.180	0.170000	0.27	0.26	0.39	2.00	0.89	0.23	0.17	0.11	6.74	0.150	0.1200	0.46	0.34	4.60	0.210	0.84	0.12	2.12	0.25	0.19	0.1300	2.84	0.15	1.25	0.132	5.34	2.07	0.12	0.17	0.18	0.25	0.17	4.17	0.14	0.49	2.15	0.45	0.42	0.16	1.06	2.79	0.170	0.17	0.41	0.53	0.85	0.61	1.13	1.47	0.190	0.43	1.67	0.11	0.163333	3.53	0.91	0.32	0.73	0.126667	0.54	0.23	0.31	1.47	0.35	0.144	0.85	0.66	0.16	0.125	0.23	0.44	0.28	14.37	0.53	0.140	0.21	0.200000	1.00	3.40	0.32	9.18	0.120	3.40	0.11	0.19	0.150000	1.19	0.45	1.74	0.25	0.19	3.71	2.27	0.36	0.18	8.24	0.21	3.67	1.51	0.59	0.12	0.161667	1.49	0.18	0.28	0.19	0.135	0.23	0.85	0.16	0.83	0.20	0.17	0.28	2.31	0.11	3.44	0.45	0.15	0.13	0.22	0.22	0.23	0.18	0.19	0.74	4.35	0.46	0.19	3.87	0.41	0.150000	0.48	2.92	0.20	0.17	0.21	0.57	0.50	0.59	4.07	0.21	1.01	0.24	1.10	0.14	0.31	0.18	0.132	1.05	0.140	0.176	0.15	4.48	0.68	1.57	4.12	1.85	0.17	3.10	0.26	0.37	0.39	0.11	0.170	1.52	0.164	0.87	0.1200	0.47	0.84	1.13	0.73	0.120	0.13	0.11	1.42	0.140000	1.14	0.56	0.200	0.12	0.79	0.25	7.81	0.160000	0.90	0.23	0.16	0.13	0.130	0.12	0.68	0.69	0.72	0.18	0.25	1.20	1.54	0.17	0.170	0.19	1.54	0.29	1.69	24.01	5.54	0.90	0.75	2.97	0.31	0.28	0.23	1.35	0.53	7.07	0.42	1.29	1.48	1.27	0.11	0.62	0.120	0.42	0.95	0.24	0.46	2.89	0.16	0.22	3.70	0.24	0.18	0.61	0.15	0.26	0.56	0.69	0.43	4.26	0.24	0.13	0.93	0.21	1.36	0.36	0.63	0.17	0.30	0.51	0.21	2.27	0.135	2.63	0.13	2.34	0.18	1.47	0.18	0.27	0.20	0.13	0.40	0.11	0.1900	0.38	0.57	0.13	5.43	0.200	0.53	0.31	0.140000	0.25	0.73	1.41	0.46	0.99	35.84	0.54	1.30	0.135	0.13	1.84	0.82	0.135	0.58	0.72	0.80	1.81	0.11	0.83	0.14	0.15	0.150	0.21	0.23	2.12	0.77	0.15	0.11	0.18	0.12	0.55	0.24	0.22	0.56	0.15	0.30	1.59	0.141429	0.134	0.31	0.25	0.68	0.60	0.37	0.13	0.12	0.34	0.21	0.60	0.14	0.22	0.45	0.36	0.16	0.15	2.41	1.18	0.40	0.36	0.20	0.136	0.66	0.11	0.95	0.16	0.72	0.21	2.45	4.14	6.04	0.47	1.78	1.28	0.23	2.30	0.90	2.42	0.15	0.15	0.25	0.13	0.31	0.47	0.14	2.50	0.17	0.16	0.29	5.12	0.53	1.05	0.47	0.85	0.120	0.17	0.21	0.33	0.24	0.36	0.13	0.24	0.90	0.74	1.31	0.58	0.12	0.86	4.48	0.15	0.54	0.34	0.32	1.39	5.21	0.160000	0.94	0.69	3.63	1.06	0.36	0.18	0.18	0.23	0.89	0.27	0.35	3.04	1.36	0.31	0.43	3.06	0.35	2.01	0.29	0.150	0.41	0.24	1.0	1
3	2020-01-04	11.380448	0.00000	0.000000	7949.995496	2.71	4.80	8.31	0.22	0.67	0.133333	0.12	3.51	0.11	9.22	0.170	0.173333	0.34	0.28	0.26	1.62	0.84	0.20	0.17	0.11	6.34	0.190	0.1225	0.39	0.32	4.81	0.185	0.85	0.14	1.72	0.21	0.17	0.1275	2.30	0.15	1.23	0.128	4.90	2.01	0.13	0.17	0.20	0.23	0.17	3.82	0.15	0.46	2.18	0.48	0.30	0.18	1.13	2.33	0.255	0.19	0.46	0.56	0.85	0.52	1.06	1.31	0.260	0.43	1.64	0.11	0.150000	3.36	1.11	0.38	0.82	0.133333	0.43	0.22	0.32	1.55	0.49	0.136	0.72	0.67	0.15	0.130	0.22	0.49	0.20	14.13	0.44	0.138	0.24	0.176667	0.95	3.80	0.21	8.22	0.119	3.71	0.11	0.19	0.143333	1.31	0.40	1.81	0.20	0.19	3.59	2.46	0.45	0.18	7.01	0.21	3.85	1.47	0.60	0.12	0.153333	1.46	0.13	0.33	0.18	0.150	0.14	0.90	0.16	0.65	0.18	0.15	0.22	1.91	0.11	3.29	0.44	0.15	0.14	0.29	0.21	0.21	0.14	0.13	0.77	4.39	0.42	0.14	3.66	0.38	0.146667	0.57	2.53	0.17	0.17	0.20	0.51	0.40	0.55	3.99	0.24	0.83	0.13	1.02	0.14	0.36	0.18	0.134	1.11	0.135	0.164	0.39	5.18	0.62	1.42	4.18	1.61	0.16	3.21	0.25	0.23	0.36	0.11	0.175	1.59	0.168	0.65	0.1175	0.68	0.76	1.01	0.69	0.170	0.13	0.11	1.02	0.133333	0.92	0.50	0.175	0.15	0.81	0.25	6.52	0.176667	0.77	0.19	0.15	0.13	0.138	0.12	0.60	0.76	0.63	0.18	0.16	1.08	1.37	0.13	0.140	0.21	1.31	0.26	1.53	22.57	5.27	0.55	0.85	2.86	0.30	0.23	0.22	1.25	0.43	7.38	0.35	1.05	1.38	1.28	0.11	0.61	0.125	0.56	0.76	0.28	0.57	2.72	0.14	0.18	3.12	0.22	0.18	0.51	0.17	0.26	0.60	0.69	0.38	4.07	0.23	0.13	0.67	0.20	1.11	0.18	0.80	0.17	0.22	0.46	0.19	1.92	0.130	2.60	0.13	2.46	0.18	1.50	0.16	0.22	0.22	0.13	0.34	0.11	0.1600	0.25	0.49	0.13	4.74	0.185	0.50	0.22	0.156667	0.25	0.53	1.48	0.46	1.01	35.48	0.50	1.37	0.150	0.14	1.81	0.62	0.130	0.54	0.80	0.74	1.72	0.11	0.81	0.14	0.15	0.140	0.21	0.23	1.81	0.93	0.18	0.11	0.18	0.13	0.61	0.18	0.15	0.46	0.23	0.30	1.44	0.142857	0.138	0.23	0.18	0.66	0.55	0.32	0.13	0.22	0.27	0.14	0.67	0.14	0.25	0.43	0.30	0.17	0.14	2.48	1.05	0.29	0.40	0.21	0.134	0.72	0.11	0.97	0.18	0.69	0.25	2.38	3.84	6.17	0.45	1.64	1.00	0.38	2.49	0.85	2.16	0.15	0.16	0.23	0.14	0.25	0.62	0.14	2.08	0.16	0.16	0.35	4.96	0.50	0.85	0.49	0.70	0.125	0.17	0.21	0.29	0.33	0.29	0.14	0.21	0.92	0.73	1.19	0.63	0.20	0.79	3.68	0.23	0.53	0.30	0.28	1.12	4.67	0.153333	1.00	0.78	3.71	1.05	0.33	0.18	0.18	0.13	0.99	0.23	0.28	3.03	1.19	0.22	0.43	3.12	0.33	2.53	0.37	0.230	0.44	0.29	1.0	1
4	2020-01-05	11.380448	0.00000	0.000000	6054.562803	2.90	4.99	8.58	0.23	0.65	0.146667	0.12	3.30	0.11	8.81	0.160	0.176667	0.38	0.29	0.38	1.57	0.88	0.19	0.16	0.11	6.72	0.195	0.1250	0.36	0.33	4.77	0.160	0.80	0.16	1.55	0.24	0.15	0.1250	2.48	0.15	1.21	0.124	5.11	1.80	0.14	0.17	0.33	0.18	0.17	3.99	0.23	0.62	2.40	0.41	0.29	0.25	1.24	3.42	0.340	0.38	0.46	0.64	0.93	0.69	0.93	1.60	0.245	0.35	1.79	0.11	0.250000	3.49	0.97	0.46	0.68	0.140000	0.40	0.22	0.28	1.55	0.37	0.128	0.73	0.55	0.14	0.135	0.25	0.41	0.21	13.39	0.42	0.136	0.24	0.153333	0.97	3.83	0.24	8.17	0.118	3.71	0.11	0.18	0.136667	1.40	0.40	1.77	0.29	0.19	4.02	2.45	0.41	0.18	7.33	0.21	3.50	1.63	0.47	0.12	0.145000	1.41	0.13	0.23	0.17	0.165	0.27	0.72	0.19	0.75	0.18	0.23	0.20	2.10	0.11	2.78	0.45	0.15	0.15	0.30	0.14	0.24	0.20	0.13	0.80	4.74	0.42	0.14	3.71	0.41	0.143333	0.56	2.56	0.19	0.17	0.16	0.65	0.55	0.63	3.83	0.31	1.04	0.16	1.16	0.14	0.38	0.18	0.136	1.01	0.130	0.152	0.31	4.90	0.61	1.46	4.65	1.56	0.15	3.04	0.20	0.23	0.39	0.11	0.180	1.41	0.172	0.60	0.1150	0.52	0.74	1.08	0.66	0.165	0.13	0.11	0.99	0.126667	1.00	0.43	0.150	0.14	0.95	0.25	6.84	0.193333	0.76	0.15	0.19	0.13	0.146	0.12	0.62	0.75	0.54	0.18	0.18	1.13	1.41	0.17	0.155	0.18	1.53	0.23	1.49	22.45	5.30	0.79	0.85	3.06	0.34	0.40	0.23	1.27	0.50	7.56	0.34	1.04	1.32	1.40	0.11	0.69	0.130	0.42	0.72	0.29	0.39	2.72	0.17	0.18	3.05	0.20	0.18	0.48	0.15	0.16	0.42	0.86	0.31	4.25	0.24	0.13	0.94	0.20	1.21	0.24	0.82	0.17	0.23	0.50	0.13	2.04	0.125	2.44	0.13	2.35	0.18	1.55	0.13	0.28	0.24	0.13	0.35	0.11	0.1525	0.18	0.49	0.13	4.91	0.170	0.51	0.29	0.173333	0.21	0.52	1.39	0.34	0.82	36.06	0.42	1.36	0.130	0.17	1.90	0.65	0.125	0.71	0.83	0.68	1.82	0.11	0.80	0.14	0.14	0.165	0.18	0.35	1.62	0.92	0.15	0.11	0.14	0.18	0.54	0.17	0.14	0.60	0.20	0.18	1.44	0.144286	0.142	0.23	0.20	0.71	0.61	0.34	0.13	0.14	0.25	0.14	0.60	0.13	0.20	0.51	0.43	0.18	0.13	2.86	1.07	0.25	0.25	0.22	0.132	0.66	0.11	0.98	0.19	0.69	0.28	2.25	3.84	6.10	0.44	1.84	0.98	0.33	2.52	0.99	2.34	0.15	0.14	0.19	0.14	0.26	0.53	0.14	2.15	0.15	0.16	0.31	5.25	0.47	0.92	0.53	0.80	0.130	0.23	0.21	0.36	0.28	0.18	0.15	0.18	0.91	0.68	1.29	0.60	0.16	0.90	3.93	0.23	0.55	0.16	0.23	1.25	4.88	0.146667	0.85	0.74	3.75	1.08	0.36	0.18	0.18	0.17	0.98	0.19	0.29	3.06	1.46	0.27	0.39	3.45	0.25	2.38	0.35	0.205	0.57	0.28	1.0	1

	JHU_cases	JHU_deaths	JHU_hospitalizations	up2date	gt_after covid vaccine	gt_side effects of vaccine	gt_effects of covid vaccine	gt_covid	gt_how long does covid last	gt_anosmia	gt_loss smell	gt_covid-19	gt_loss taste	gt_loss of smell	gt_chest pain	gt_covid symptoms	gt_sars-cov 2	gt_chest tightness	gt_covid nhs	gt_quarantine	gt_covid-19 who	gt_sars-cov-2	gt_feeling exhausted	gt_nose bleed	gt_feeling tired	gt_joints aching	gt_fever	gt2_Abdominal obesity	gt2_Abdominal pain	gt2_Acne	gt2_Actinic keratosis	gt2_Acute bronchitis	gt2_Adrenal crisis	gt2_Ageusia	gt2_Alcoholism	gt2_Allergic conjunctivitis	gt2_Allergy	gt2_Amblyopia	gt2_Amenorrhea	gt2_Amnesia	gt2_Anal fissure	gt2_Anaphylaxis	gt2_Anemia	gt2_Angina pectoris	gt2_Angioedema	gt2_Angular cheilitis	gt2_Anosmia	gt2_Anxiety	gt2_Aphasia	gt2_Aphonia	gt2_Apnea	gt2_Arthralgia	gt2_Arthritis	gt2_Ascites	gt2_Asperger syndrome	gt2_Asphyxia	gt2_Asthma	gt2_Astigmatism	gt2_Ataxia	gt2_Atheroma	gt2_Attention deficit hyperactivity disorder	gt2_Auditory hallucination	gt2_Autoimmune disease	gt2_Avoidant personality disorder	gt2_Back pain	gt2_Bacterial vaginosis	gt2_Balance disorder	gt2_Beau's lines	gt2_Bell's palsy	gt2_Biliary colic	gt2_Binge eating	gt2_Bleeding	gt2_Bleeding on probing	gt2_Blepharospasm	gt2_Bloating	gt2_Blood in stool	gt2_Blurred vision	gt2_Blushing	gt2_Boil	gt2_Bone fracture	gt2_Bone tumor	gt2_Bowel obstruction	gt2_Bradycardia	gt2_Braxton Hicks contractions	gt2_Breakthrough bleeding	gt2_Breast pain	gt2_Bronchitis	gt2_Bruise	gt2_Bruxism	gt2_Bunion	gt2_Burn	gt2_Burning Chest Pain	gt2_Burning mouth syndrome	gt2_Candidiasis	gt2_Canker sore	gt2_Cardiac arrest	gt2_Carpal tunnel syndrome	gt2_Cataplexy	gt2_Cataract	gt2_Chancre	gt2_Cheilitis	gt2_Chest pain	gt2_Chills	gt2_Chorea	gt2_Chronic pain	gt2_Cirrhosis	gt2_Cleft lip and cleft palate	gt2_Clouding of consciousness	gt2_Cluster headache	gt2_Colitis	gt2_Coma	gt2_Common cold	gt2_Compulsive behavior	gt2_Compulsive hoarding	gt2_Confusion	gt2_Congenital heart defect	gt2_Conjunctivitis	gt2_Constipation	gt2_Convulsion	gt2_Cough	gt2_Crackles	gt2_Cramp	gt2_Crepitus	gt2_Croup	gt2_Cyanosis	gt2_Dandruff	gt2_Delayed onset muscle soreness	gt2_Dementia	gt2_Dentin hypersensitivity	gt2_Depersonalization	gt2_Depression	gt2_Dermatitis	gt2_Desquamation	gt2_Developmental disability	gt2_Diabetes	gt2_Diabetic ketoacidosis	gt2_Diarrhea	gt2_Dizziness	gt2_Dry eye syndrome	gt2_Dysautonomia	gt2_Dysgeusia	gt2_Dysmenorrhea	gt2_Dyspareunia	gt2_Dysphagia	gt2_Dysphoria	gt2_Dystonia	gt2_Dysuria	gt2_Ear pain	gt2_Eczema	gt2_Edema	gt2_Encephalitis	gt2_Encephalopathy	gt2_Epidermoid cyst	gt2_Epilepsy	gt2_Epiphora	gt2_Erectile dysfunction	gt2_Erythema	gt2_Erythema chronicum migrans	gt2_Esophagitis	gt2_Excessive daytime sleepiness	gt2_Eye pain	gt2_Eye strain	gt2_Facial nerve paralysis	gt2_Facial swelling	gt2_Fasciculation	gt2_Fatigue	gt2_Fatty liver disease	gt2_Fecal incontinence	gt2_Fever	gt2_Fibrillation	gt2_Fibrocystic breast changes	gt2_Fibromyalgia	gt2_Flatulence	gt2_Floater	gt2_Focal seizure	gt2_Folate deficiency	gt2_Food craving	gt2_Food intolerance	gt2_Frequent urination	gt2_Gastroesophageal reflux disease	gt2_Gastroparesis	gt2_Generalized anxiety disorder	gt2_Generalized tonic–clonic seizure	gt2_Genital wart	gt2_Gingival recession	gt2_Gingivitis	gt2_Globus pharyngis	gt2_Goitre	gt2_Gout	gt2_Grandiosity	gt2_Granuloma	gt2_Guilt	gt2_Hair loss	gt2_Halitosis	gt2_Hay fever	gt2_Headache	gt2_Heart arrhythmia	gt2_Heart murmur	gt2_Heartburn	gt2_Hematochezia	gt2_Hematoma	gt2_Hematuria	gt2_Hemolysis	gt2_Hemoptysis	gt2_Hemorrhoids	gt2_Hepatic encephalopathy	gt2_Hepatitis	gt2_Hepatotoxicity	gt2_Hiccup	gt2_Hip pain	gt2_Hives	gt2_Hot flash	gt2_Hydrocephalus	gt2_Hypercalcaemia	gt2_Hypercapnia	gt2_Hypercholesterolemia	gt2_Hyperemesis gravidarum	gt2_Hyperglycemia	gt2_Hyperkalemia	gt2_Hyperlipidemia	gt2_Hypermobility	gt2_Hyperpigmentation	gt2_Hypersomnia	gt2_Hypertension	gt2_Hyperthermia	gt2_Hyperthyroidism	gt2_Hypertriglyceridemia	gt2_Hypertrophy	gt2_Hyperventilation	gt2_Hypocalcaemia	gt2_Hypochondriasis	gt2_Hypoglycemia	gt2_Hypogonadism	gt2_Hypokalemia	gt2_Hypomania	gt2_Hyponatremia	gt2_Hypotension	gt2_Hypothyroidism	gt2_Hypoxemia	gt2_Hypoxia	gt2_Impetigo	gt2_Implantation bleeding	gt2_Impulsivity	gt2_Indigestion	gt2_Infection	gt2_Inflammation	gt2_Inflammatory bowel disease	gt2_Ingrown hair	gt2_Insomnia	gt2_Insulin resistance	gt2_Intermenstrual bleeding	gt2_Intracranial pressure	gt2_Iron deficiency	gt2_Irregular menstruation	gt2_Itch	gt2_Jaundice	gt2_Kidney failure	gt2_Kidney stone	gt2_Knee Pain	gt2_Kyphosis	gt2_Lactose intolerance	gt2_Laryngitis	gt2_Leg cramps	gt2_Lesion	gt2_Leukorrhea	gt2_Lightheadedness	gt2_Low back pain	gt2_Low-grade fever	gt2_Lymphedema	gt2_Major depressive disorder	gt2_Malabsorption	gt2_Male infertility	gt2_Manic Disorder	gt2_Melasma	gt2_Melena	gt2_Meningitis	gt2_Menorrhagia	gt2_Middle back pain	gt2_Migraine	gt2_Milium	gt2_Mitral insufficiency	gt2_Mood disorder	gt2_Mood swing	gt2_Morning sickness	gt2_Motion sickness	gt2_Mouth ulcer	gt2_Muscle atrophy	gt2_Muscle weakness	gt2_Myalgia	gt2_Mydriasis	gt2_Myocardial infarction	gt2_Myoclonus	gt2_Nasal congestion	gt2_Nasal polyp	gt2_Nausea	gt2_Neck mass	gt2_Neck pain	gt2_Neonatal jaundice	gt2_Nerve injury	gt2_Neuralgia	gt2_Neutropenia	gt2_Night sweats	gt2_Night terror	gt2_Nocturnal enuresis	gt2_Nodule	gt2_Nosebleed	gt2_Nystagmus	gt2_Obesity	gt2_Onychorrhexis	gt2_Oral candidiasis	gt2_Orthostatic hypotension	gt2_Osteopenia	gt2_Osteophyte	gt2_Osteoporosis	gt2_Otitis	gt2_Otitis externa	gt2_Otitis media	gt2_Pain	gt2_Palpitations	gt2_Panic attack	gt2_Papule	gt2_Paranoia	gt2_Paresthesia	gt2_Pelvic inflammatory disease	gt2_Pericarditis	gt2_Periodontal disease	gt2_Periorbital puffiness	gt2_Peripheral neuropathy	gt2_Perspiration	gt2_Petechia	gt2_Phlegm	gt2_Photodermatitis	gt2_Photophobia	gt2_Photopsia	gt2_Pleural effusion	gt2_Pleurisy	gt2_Pneumonia	gt2_Podalgia	gt2_Polycythemia	gt2_Polydipsia	gt2_Polyneuropathy	gt2_Polyuria	gt2_Poor posture	gt2_Post-nasal drip	gt2_Postural orthostatic tachycardia syndrome	gt2_Prediabetes	gt2_Proteinuria	gt2_Pruritus ani	gt2_Psychosis	gt2_Ptosis	gt2_Pulmonary edema	gt2_Pulmonary hypertension	gt2_Purpura	gt2_Pus	gt2_Pyelonephritis	gt2_Radiculopathy	gt2_Rectal pain	gt2_Rectal prolapse	gt2_Red eye	gt2_Renal colic	gt2_Restless legs syndrome	gt2_Rheum	gt2_Rhinitis	gt2_Rhinorrhea	gt2_Rosacea	gt2_Round ligament pain	gt2_Rumination	gt2_Scar	gt2_Sciatica	gt2_Scoliosis	gt2_Seborrheic dermatitis	gt2_Self-harm	gt2_Sensitivity to sound	gt2_Sexual dysfunction	gt2_Shallow breathing	gt2_Sharp pain	gt2_Shivering	gt2_Shortness of breath	gt2_Shyness	gt2_Sinusitis	gt2_Skin condition	gt2_Skin rash	gt2_Skin tag	gt2_Skin ulcer	gt2_Sleep apnea	gt2_Sleep deprivation	gt2_Sleep disorder	gt2_Snoring	gt2_Sore throat	gt2_Spasticity	gt2_Splenomegaly	gt2_Sputum	gt2_Stomach rumble	gt2_Strabismus	gt2_Stretch marks	gt2_Stridor	gt2_Stroke	gt2_Stuttering	gt2_Subdural hematoma	gt2_Suicidal ideation	gt2_Swelling	gt2_Swollen feet	gt2_Swollen lymph nodes	gt2_Syncope	gt2_Tachycardia	gt2_Tachypnea	gt2_Telangiectasia	gt2_Tenderness	gt2_Testicular pain	gt2_Throat irritation	gt2_Thrombocytopenia	gt2_Thyroid nodule	gt2_Tic	gt2_Tinnitus	gt2_Tonsillitis	gt2_Toothache	gt2_Tremor	gt2_Trichoptilosis	gt2_Tumor	gt2_Type 2 diabetes	gt2_Unconsciousness	gt2_Underweight	gt2_Upper respiratory tract infection	gt2_Urethritis	gt2_Urinary incontinence	gt2_Urinary tract infection	gt2_Urinary urgency	gt2_Uterine contraction	gt2_Vaginal bleeding	gt2_Vaginal discharge	gt2_Vaginitis	gt2_Varicose veins	gt2_Vasculitis	gt2_Ventricular fibrillation	gt2_Ventricular tachycardia	gt2_Vertigo	gt2_Viral pneumonia	gt2_Visual acuity	gt2_Vomiting	gt2_Wart	gt2_Water retention	gt2_Weakness	gt2_Weight gain	gt2_Wheeze	gt2_Xeroderma	gt2_Xerostomia	gt2_Yawn	gt2_hyperhidrosis	gt2_pancreatitis	neighbor_Alabama	neighbor_Alaska	neighbor_Arizona	neighbor_Arkansas	neighbor_California	neighbor_Colorado	neighbor_Connecticut	neighbor_Delaware	neighbor_Florida	neighbor_Georgia	neighbor_Hawaii	neighbor_Idaho	neighbor_Illinois	neighbor_Indiana	neighbor_Iowa	neighbor_Kansas	neighbor_Kentucky	neighbor_Louisiana	neighbor_Maine	neighbor_Maryland	neighbor_Massachusetts	neighbor_Michigan	neighbor_Minnesota	neighbor_Missouri	neighbor_Montana	neighbor_Nebraska	neighbor_Nevada	neighbor_New Hampshire	neighbor_New Jersey	neighbor_New Mexico	neighbor_New York	neighbor_North Carolina	neighbor_North Dakota	neighbor_Ohio	neighbor_Oklahoma	neighbor_Oregon	neighbor_Pennsylvania	neighbor_Rhode Island	neighbor_South Carolina	neighbor_South Dakota	neighbor_Tennessee	neighbor_Texas	neighbor_Utah	neighbor_Vermont	neighbor_Virginia	neighbor_Washington	neighbor_West Virginia	neighbor_Wisconsin	neighbor_Wyoming	neighbor_Mississippi	State_Name_Alabama	State_Name_Alaska	State_Name_Arizona	State_Name_Arkansas	State_Name_California	State_Name_Colorado	State_Name_Connecticut	State_Name_Delaware	State_Name_Florida	State_Name_Georgia	State_Name_Hawaii	State_Name_Idaho	State_Name_Illinois	State_Name_Indiana	State_Name_Iowa	State_Name_Kansas	State_Name_Kentucky	State_Name_Louisiana	State_Name_Maine	State_Name_Maryland	State_Name_Massachusetts	State_Name_Michigan	State_Name_Minnesota	State_Name_Mississippi	State_Name_Missouri	State_Name_Montana	State_Name_Nebraska	State_Name_Nevada	State_Name_New Hampshire	State_Name_New Jersey	State_Name_New Mexico	State_Name_New York	State_Name_North Carolina	State_Name_North Dakota	State_Name_Ohio	State_Name_Oklahoma	State_Name_Oregon	State_Name_Pennsylvania	State_Name_Rhode Island	State_Name_South Carolina	State_Name_South Dakota	State_Name_Tennessee	State_Name_Texas	State_Name_Utah	State_Name_Vermont	State_Name_Virginia	State_Name_Washington	State_Name_West Virginia	State_Name_Wisconsin	State_Name_Wyoming
count	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.00000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.00000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.00000	37201.000000	37201.000000	37201.000000	37201.000000	37201.00000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.00000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000	37201.000000
mean	1642.996694	22.461412	129.596933	239.810169	966.864861	300.261378	215.281892	86737.161589	289.107101	27.279207	156.702446	5487.864721	176.667034	148.596383	1137.662282	4830.924348	0.423490	117.438872	4.515029	2952.605131	97.357118	82.490776	1.223157	136.716192	28.221853	2.509515	3565.609128	2.224520	4.211325	7.649605	0.327083	0.381364	0.109883	0.219468	3.931982	0.064682	11.162166	0.213526	0.153194	0.405745	0.269757	0.459716	2.228877	0.853727	0.214926	0.060918	0.242004	7.342641	0.263803	0.046043	0.455591	0.328220	4.825432	0.223287	1.108655	0.171113	2.260597	0.316388	0.215155	0.155737	3.380084	0.082201	1.448329	0.229761	5.082650	1.421830	0.151113	0.104212	0.265999	0.238230	0.251991	3.769222	0.163597	0.433256	2.019202	0.455902	0.393479	0.211680	0.945504	2.999670	0.219953	0.290896	0.583571	0.421845	0.677466	0.469020	0.606557	1.537552	0.308325	0.423398	1.857737	0.067650	0.177879	2.685605	0.878447	0.471364	0.811826	0.104382	0.521374	0.200481	0.261375	1.376410	0.308484	0.162013	0.925049	0.698840	0.167758	0.089006	0.242922	0.490642	0.308703	10.943745	0.615894	0.167009	0.265562	0.277838	0.826816	3.317770	0.317632	5.494946	0.059775	2.938066	0.083635	0.244811	0.100265	1.066879	0.444750	2.093487	0.280338	0.090839	4.320914	2.277043	0.370444	0.151334	7.993027	0.283856	3.860419	1.537984	0.651461	0.075846	0.162223	1.187917	0.102158	0.350593	0.214050	0.177243	0.172793	0.716377	0.200848	0.907026	0.214110	0.246937	0.253640	2.257320	0.091179	2.643434	0.493368	0.257805	0.112342	0.285428	0.177484	0.28177	0.247987	0.156521	0.757898	4.347871	0.383482	0.181122	4.242673	0.516230	0.109744	0.539432	2.069475	0.244705	0.109084	0.224942	0.440011	0.586238	0.491773	3.779818	0.166749	1.024028	0.209863	0.863294	0.108243	0.416217	0.108867	0.162508	0.956812	0.175044	0.186710	0.324563	4.445122	0.515993	1.783457	4.138789	1.998067	0.193946	2.782231	0.221360	0.387623	0.381611	0.067587	0.098642	1.402507	0.097531	0.879296	0.104956	0.508539	0.830373	1.107517	0.640277	0.185522	0.114495	0.088974	1.266390	0.106182	1.121007	0.579709	0.200445	0.210298	0.839792	0.089375	6.981148	0.355204	1.012353	0.262640	0.290627	0.139757	0.183604	0.146041	0.721842	0.641578	0.717592	0.146621	0.328005	1.221746	1.562581	0.410885	0.238816	0.203684	1.076106	0.441314	1.453992	23.547391	5.459019	1.005626	0.709681	2.986322	0.391171	0.225312	0.323016	1.556976	0.437138	7.230728	0.490662	1.403749	1.304477	1.531915	0.153611	0.667295	0.069680	0.421260	1.052838	0.201736	0.484309	2.756755	0.266084	0.265207	3.960642	0.154553	0.165344	0.591108	0.253471	0.289800	0.449997	0.675633	0.330527	4.086483	0.238928	0.186927	0.997439	0.238305	0.894380	0.303466	0.737719	0.155804	0.272075	0.517059	0.241671	2.24946	0.071053	1.886288	0.092292	2.111539	0.142397	1.445533	0.217087	0.288937	0.238081	0.173985	0.281796	0.096067	0.15003	0.379989	0.484197	0.098340	5.122530	0.13753	0.497266	0.281156	0.207329	0.166162	0.822688	1.127099	0.426365	0.779262	34.302607	0.439723	1.46434	0.163608	0.175485	1.849272	0.565777	0.103019	0.679260	0.704358	0.814511	1.788223	0.141154	0.665511	0.093213	0.092147	0.179131	0.273304	0.184713	1.319595	0.889690	0.193660	0.065331	0.203206	0.160613	0.520711	0.220946	0.260138	0.582827	0.295544	0.269961	1.647219	0.102412	0.076359	0.378657	0.232703	0.590554	0.472933	0.412738	0.074451	0.149371	0.345953	0.173269	0.538049	0.090880	0.305547	0.408080	0.353565	0.179154	0.088123	2.460581	1.067069	0.290785	0.334778	0.245754	0.106790	0.601926	0.058222	0.778080	0.219264	0.800860	0.314626	1.953237	3.895940	6.334318	0.512038	1.623754	1.176944	0.345340	2.184036	0.833432	1.887773	0.095309	0.155792	0.182519	0.104091	0.350930	0.499001	0.073749	2.678848	0.223538	0.113419	0.381012	5.401614	0.581137	0.945389	0.676490	0.917601	0.095219	0.258673	0.164246	0.253894	0.257782	0.452813	0.168641	0.297487	0.861295	0.644043	0.958067	0.606450	0.148172	1.078524	4.014653	0.232984	0.588387	0.226964	0.204251	1.307818	4.053234	0.067888	0.817711	0.543586	2.723427	0.761255	0.471089	0.188512	0.067815	0.258866	1.053671	0.060731	0.223656	2.736969	1.227934	0.293425	0.426625	3.109280	0.287278	2.292905	0.443379	0.179036	0.488471	0.368599	1327.719470	217.797828	2034.353942	861.475552	8179.667966	1303.218758	755.674283	291.037822	6278.364049	2540.742399	210.420096	433.860783	3138.071907	1751.074622	878.234026	745.770490	1206.369963	1196.553103	209.977205	1157.366254	1685.881347	2444.214994	1401.553318	1537.577269	267.204833	503.653316	671.060670	282.627564	2607.112363	486.184162	5882.125884	2392.479234	245.316094	2979.518588	971.534556	605.502540	3173.699659	350.453859	1380.724927	249.486277	1927.012822	6913.934222	898.356603	117.978092	1608.310637	1226.405419	496.411172	1562.249241	157.871912	772.581463	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.020026	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999	0.019999
std	4589.492744	58.320425	222.529119	184.199940	1860.902082	860.911626	692.092902	49879.919783	769.121134	206.544173	533.877253	7599.045473	590.841522	526.643874	1434.119878	4401.861800	11.716711	463.056179	69.617293	4111.938931	424.346806	370.925946	20.343860	484.900113	196.304827	33.331048	2799.152905	0.545928	0.518877	1.206582	0.089734	0.203964	0.067302	0.165150	0.644033	0.055037	2.049919	0.078566	0.039093	0.097349	0.081695	0.118838	0.493838	0.223363	0.083358	0.045597	0.184194	0.969800	0.102122	0.037843	0.120233	0.087179	0.981536	0.085675	0.441319	0.270638	0.571206	0.122133	0.090843	0.042588	0.818402	0.056224	0.277393	0.089012	0.629778	0.399944	0.038505	0.058974	0.211819	0.088582	0.086206	0.518650	0.044619	0.141227	0.288387	0.108156	0.109454	0.079437	0.189443	0.658813	0.094405	0.093934	0.119415	0.089742	0.141889	0.087828	0.291549	0.284760	0.080980	0.085769	0.383977	0.051675	0.052636	0.468132	0.119544	0.125282	0.128367	0.064962	0.155910	0.062259	0.100070	0.219517	0.122396	0.087060	0.183228	0.159627	0.062237	0.054594	0.085614	0.121315	0.147525	6.132593	0.115714	0.087453	0.079285	0.108800	0.197069	0.445586	0.087962	2.395301	0.045652	0.413972	0.056821	0.136911	0.059914	0.202589	0.104236	0.569267	0.077982	0.060346	0.661276	0.351888	0.099583	0.051693	1.502710	0.093963	0.529150	0.219759	0.125073	0.054346	0.050347	0.187742	0.056518	0.102488	0.083446	0.060737	0.047565	0.184353	0.083168	0.230056	0.111483	0.111505	0.130423	0.526821	0.061073	0.462681	0.096964	0.481927	0.059824	0.085290	0.045286	0.10082	0.205409	0.045594	0.164024	0.548570	0.101449	0.056185	1.745725	0.151499	0.063587	0.107715	0.348704	0.081021	0.058960	0.077896	0.089439	0.117073	0.094690	0.540364	0.049454	0.196645	0.072436	0.153480	0.057488	0.088588	0.063770	0.049911	0.176957	0.052994	0.066662	0.094550	0.694195	0.101951	0.598155	0.516099	0.434475	0.066200	0.464012	0.084227	0.144072	0.092662	0.061485	0.063644	0.188999	0.070156	0.229291	0.061377	0.095377	0.141488	0.214752	0.125095	0.066630	0.064237	0.063064	0.368522	0.061229	0.219211	0.131392	0.070460	0.095643	0.221749	0.054334	1.596895	0.167172	0.222508	0.097705	0.104193	0.041719	0.061611	0.042554	0.150322	0.149327	0.164113	0.036447	0.103147	0.233984	0.298250	0.177780	0.094049	0.082083	0.208902	0.113871	0.223313	8.411747	0.786016	0.164177	0.164057	0.414602	0.094522	0.075956	0.180643	0.298121	0.091191	1.661685	0.118604	0.362461	0.208900	0.263665	0.043619	0.123194	0.057358	0.107694	0.272461	0.062724	0.089234	0.358501	0.140193	0.183888	0.758183	0.061734	0.051479	0.112011	0.096328	0.087105	0.154127	0.179805	0.079512	0.512951	0.081385	0.100023	0.182619	0.084688	0.157096	0.103269	0.108844	0.042120	0.088367	0.108615	0.089315	0.36011	0.052141	0.732544	0.055112	0.345236	0.035824	0.192274	0.090983	0.082669	0.087921	0.054613	0.091074	0.057460	0.04196	0.123000	0.123626	0.058247	0.929657	0.03579	0.091092	0.081470	0.072335	0.046897	0.232267	0.252951	0.152700	0.200095	3.681106	0.083371	0.20772	0.042887	0.062289	0.236163	0.165099	0.060938	0.139605	0.152195	0.169442	0.376893	0.043107	0.263539	0.116521	0.058767	0.064853	0.098330	0.075136	0.560540	0.155622	0.077965	0.058938	0.096430	0.046269	0.115407	0.096079	0.111067	0.103458	0.106173	0.099549	0.278741	0.056736	0.051770	0.131030	0.081638	0.101374	0.096368	0.085382	0.054467	0.044367	0.083111	0.065222	0.101837	0.059609	0.093993	0.137756	0.086957	0.058963	0.055978	0.417083	0.158790	0.090534	0.078794	0.083219	0.057841	0.123102	0.045726	0.116799	0.101946	0.212136	0.089945	0.582688	0.606909	1.484545	0.108460	0.220600	0.266739	0.092933	0.302573	0.141953	0.646125	0.063911	0.043921	0.064398	0.061922	0.103055	0.132451	0.056323	0.527234	0.123687	0.067366	0.101521	0.747575	0.167573	0.157793	0.127647	0.191350	0.061198	0.086242	0.045334	0.083751	0.110987	0.152294	0.054106	0.085000	0.214930	0.145842	0.164854	0.117778	0.039428	0.309929	0.826865	0.086244	0.136575	0.124381	0.070935	0.228164	0.689213	0.053341	0.139305	0.106252	0.528583	0.160877	0.112833	0.076779	0.056125	0.097926	0.169899	0.053682	0.081589	0.481990	0.203958	0.087904	0.097998	0.519479	0.090348	0.555266	0.097567	0.056662	0.130634	0.091791	1777.121206	400.107915	2668.066812	1181.408242	12606.656158	1842.571852	1682.655019	474.923196	14571.299146	4288.389593	547.087449	500.766977	5717.612569	2542.248441	1549.635658	1618.267099	1667.903688	2258.386126	318.025668	2357.062860	3151.439206	5052.699771	2025.565127	2200.359874	372.512287	746.204161	871.662002	552.809509	4900.333245	746.226191	12044.831200	3907.653535	335.826951	4076.669346	1396.863726	968.288348	4538.722796	910.759634	2454.836057	371.977558	3433.185678	10186.784345	1258.928274	309.490778	2625.894387	2178.501511	747.886818	2253.383594	244.802494	1275.577884	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140092	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000	0.140000
min	-9005.000000	-364.000000	0.000000	0.305869	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.570000	1.660000	3.090000	0.100000	0.050000	0.000000	0.000000	1.690000	0.000000	4.260000	0.050000	0.040000	0.130000	0.090000	0.130000	0.630000	0.240000	0.060000	0.000000	0.000000	3.550000	0.040000	0.000000	0.140000	0.100000	2.120000	0.040000	0.430000	0.040000	0.620000	0.080000	0.040000	0.040000	1.130000	0.000000	0.410000	0.070000	2.100000	0.360000	0.040000	0.000000	0.060000	0.060000	0.090000	1.460000	0.040000	0.100000	0.760000	0.120000	0.110000	0.070000	0.310000	1.160000	0.040000	0.080000	0.140000	0.110000	0.120000	0.130000	0.100000	0.570000	0.090000	0.110000	0.700000	0.000000	0.040000	0.960000	0.270000	0.110000	0.210000	0.000000	0.070000	0.050000	0.070000	0.530000	0.060000	0.040000	0.300000	0.200000	0.050000	0.000000	0.080000	0.130000	0.090000	2.400000	0.230000	0.040000	0.080000	0.060000	0.170000	1.190000	0.100000	1.290000	0.000000	1.070000	0.000000	0.030000	0.000000	0.300000	0.110000	0.690000	0.080000	0.000000	1.800000	0.890000	0.100000	0.030000	2.900000	0.060000	1.790000	0.610000	0.180000	0.000000	0.040000	0.330000	0.000000	0.080000	0.060000	0.040000	0.050000	0.210000	0.060000	0.270000	0.050000	0.050000	0.070000	0.780000	0.000000	0.950000	0.150000	0.050000	0.000000	0.100000	0.050000	0.09000	0.050000	0.040000	0.200000	1.980000	0.100000	0.050000	1.600000	0.110000	0.000000	0.160000	0.960000	0.050000	0.000000	0.040000	0.120000	0.200000	0.140000	1.760000	0.040000	0.370000	0.060000	0.240000	0.000000	0.120000	0.000000	0.040000	0.330000	0.050000	0.040000	0.110000	1.290000	0.170000	0.460000	1.720000	0.640000	0.040000	1.280000	0.070000	0.100000	0.110000	0.000000	0.000000	0.590000	0.000000	0.250000	0.000000	0.150000	0.300000	0.300000	0.230000	0.040000	0.000000	0.000000	0.160000	0.000000	0.400000	0.150000	0.030000	0.050000	0.160000	0.000000	2.270000	0.090000	0.330000	0.040000	0.070000	0.040000	0.040000	0.040000	0.210000	0.130000	0.200000	0.040000	0.060000	0.360000	0.570000	0.080000	0.060000	0.040000	0.310000	0.120000	0.560000	8.830000	2.440000	0.330000	0.150000	1.240000	0.110000	0.050000	0.060000	0.520000	0.120000	2.630000	0.120000	0.370000	0.470000	0.520000	0.050000	0.280000	0.000000	0.110000	0.290000	0.040000	0.170000	1.070000	0.040000	0.050000	1.460000	0.050000	0.040000	0.140000	0.040000	0.080000	0.100000	0.140000	0.110000	1.770000	0.080000	0.030000	0.380000	0.070000	0.280000	0.080000	0.290000	0.050000	0.080000	0.190000	0.050000	0.93000	0.000000	0.300000	0.000000	0.770000	0.040000	0.560000	0.070000	0.060000	0.050000	0.040000	0.080000	0.000000	0.04000	0.070000	0.100000	0.000000	1.650000	0.04000	0.120000	0.090000	0.040000	0.040000	0.170000	0.360000	0.100000	0.210000	16.280000	0.140000	0.65000	0.050000	0.060000	0.730000	0.140000	0.000000	0.140000	0.210000	0.210000	0.690000	0.040000	0.120000	0.000000	0.000000	0.040000	0.050000	0.050000	0.230000	0.230000	0.040000	0.000000	0.040000	0.040000	0.170000	0.040000	0.060000	0.160000	0.060000	0.070000	0.590000	0.000000	0.000000	0.070000	0.060000	0.120000	0.120000	0.130000	0.000000	0.040000	0.100000	0.050000	0.170000	0.000000	0.070000	0.080000	0.100000	0.050000	0.000000	0.840000	0.420000	0.090000	0.100000	0.100000	0.000000	0.170000	0.000000	0.280000	0.060000	0.250000	0.100000	0.430000	1.640000	2.430000	0.140000	0.660000	0.390000	0.100000	0.900000	0.280000	0.460000	0.000000	0.040000	0.040000	0.000000	0.100000	0.110000	0.000000	0.880000	0.070000	0.000000	0.130000	2.190000	0.120000	0.320000	0.210000	0.310000	0.000000	0.070000	0.040000	0.080000	0.060000	0.100000	0.040000	0.110000	0.250000	0.180000	0.360000	0.220000	0.040000	0.340000	1.330000	0.070000	0.150000	0.040000	0.040000	0.470000	1.490000	0.000000	0.270000	0.110000	0.910000	0.180000	0.130000	0.050000	0.000000	0.050000	0.400000	0.000000	0.050000	0.970000	0.470000	0.080000	0.120000	1.000000	0.090000	0.710000	0.130000	0.050000	0.140000	0.100000	0.000000	0.000000	0.000000	-400.000000	-3935.000000	-2.000000	-15.000000	-27.000000	-1448.000000	0.000000	-79.000000	-2.000000	0.000000	0.000000	-947.000000	-168.000000	-3.000000	-119.000000	-7.000000	-3.000000	-280.000000	0.000000	0.000000	-8492.000000	-1.000000	-43.000000	-51.000000	-63.000000	-9005.000000	0.000000	0.000000	0.000000	-35.000000	0.000000	-1294.000000	0.000000	0.000000	-1260.000000	-22.000000	-4.000000	-42.000000	0.000000	0.000000	-1.000000	-232.000000	-39.000000	-4.000000	0.000000	-8.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
25%	39.000000	0.000000	16.000000	117.239542	0.000000	0.000000	0.000000	54508.215741	0.000000	0.000000	0.000000	1427.348263	0.000000	0.000000	0.000000	1653.298169	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	1885.633898	1.800000	3.950000	6.890000	0.260000	0.230000	0.090000	0.090000	3.610000	0.040000	9.860000	0.170000	0.130000	0.340000	0.220000	0.390000	1.850000	0.690000	0.170000	0.040000	0.100000	6.770000	0.190000	0.000000	0.370000	0.280000	4.360000	0.170000	0.960000	0.120000	1.810000	0.250000	0.160000	0.130000	2.740000	0.060000	1.250000	0.170000	4.740000	1.090000	0.130000	0.080000	0.180000	0.180000	0.200000	3.470000	0.140000	0.310000	1.830000	0.400000	0.330000	0.170000	0.820000	2.590000	0.170000	0.230000	0.500000	0.360000	0.590000	0.420000	0.390000	1.370000	0.260000	0.370000	1.620000	0.042500	0.140000	2.380000	0.810000	0.390000	0.730000	0.070000	0.400000	0.155000	0.200000	1.250000	0.230000	0.110000	0.790000	0.580000	0.130000	0.070000	0.200000	0.410000	0.240000	7.500000	0.550000	0.125000	0.220000	0.200000	0.680000	3.080000	0.260000	3.810000	0.040000	2.710000	0.060000	0.140000	0.080000	0.940000	0.380000	1.810000	0.240000	0.070000	3.920000	2.060000	0.300000	0.120000	6.960000	0.220000	3.540000	1.410000	0.560000	0.057500	0.130000	1.080000	0.080000	0.280000	0.170000	0.140000	0.140000	0.590000	0.150000	0.760000	0.152500	0.180000	0.200000	1.850000	0.070000	2.330000	0.430000	0.140000	0.090000	0.230000	0.150000	0.22000	0.160000	0.120000	0.630000	4.030000	0.310000	0.140000	3.380000	0.400000	0.080000	0.470000	1.910000	0.200000	0.080000	0.170000	0.380000	0.510000	0.430000	3.500000	0.130000	0.890000	0.160000	0.770000	0.090000	0.360000	0.090000	0.130000	0.850000	0.140000	0.140000	0.270000	4.080000	0.450000	1.450000	3.860000	1.680000	0.150000	2.560000	0.175000	0.310000	0.320000	0.040000	0.070000	1.290000	0.070000	0.710000	0.080000	0.450000	0.750000	0.970000	0.560000	0.140000	0.090000	0.060000	0.970000	0.090000	0.970000	0.480000	0.150000	0.160000	0.690000	0.070000	5.860000	0.250000	0.850000	0.195000	0.220000	0.110000	0.140000	0.120000	0.610000	0.550000	0.600000	0.120000	0.250000	1.070000	1.340000	0.290000	0.180000	0.150000	0.940000	0.360000	1.330000	19.790000	4.940000	0.900000	0.590000	2.710000	0.320000	0.180000	0.240000	1.330000	0.380000	6.030000	0.420000	1.130000	1.170000	1.370000	0.125000	0.590000	0.040000	0.340000	0.840000	0.155000	0.430000	2.550000	0.180000	0.190000	3.370000	0.120000	0.130000	0.520000	0.190000	0.230000	0.360000	0.580000	0.290000	3.790000	0.190000	0.130000	0.870000	0.190000	0.790000	0.230000	0.670000	0.130000	0.210000	0.460000	0.180000	2.02000	0.050000	1.370000	0.080000	1.880000	0.120000	1.320000	0.170000	0.240000	0.180000	0.140000	0.230000	0.070000	0.12000	0.280000	0.400000	0.070000	4.490000	0.11750	0.430000	0.230000	0.150000	0.140000	0.630000	0.940000	0.310000	0.630000	32.380000	0.390000	1.34000	0.140000	0.140000	1.720000	0.480000	0.080000	0.580000	0.600000	0.690000	1.530000	0.110000	0.490000	0.020000	0.070000	0.140000	0.210000	0.140000	0.940000	0.800000	0.150000	0.040000	0.150000	0.130000	0.450000	0.160000	0.170000	0.520000	0.220000	0.210000	1.450000	0.080000	0.054286	0.280000	0.180000	0.530000	0.400000	0.360000	0.050000	0.120000	0.290000	0.140000	0.470000	0.070000	0.240000	0.320000	0.300000	0.140000	0.070000	2.190000	0.960000	0.230000	0.290000	0.200000	0.090000	0.530000	0.040000	0.710000	0.170000	0.690000	0.260000	1.530000	3.510000	5.240000	0.440000	1.500000	0.980000	0.290000	2.010000	0.740000	1.440000	0.060000	0.130000	0.133333	0.080000	0.280000	0.410000	0.050000	2.310000	0.170000	0.090000	0.320000	4.950000	0.460000	0.840000	0.590000	0.780000	0.070000	0.200000	0.130000	0.200000	0.180000	0.340000	0.130000	0.250000	0.760000	0.550000	0.860000	0.540000	0.120000	0.840000	3.420000	0.180000	0.490000	0.146000	0.150000	1.160000	3.600000	0.040000	0.720000	0.480000	2.350000	0.660000	0.390000	0.140000	0.040000	0.190000	0.950000	0.000000	0.180000	2.380000	1.100000	0.240000	0.360000	2.770000	0.230000	1.930000	0.380000	0.140000	0.410000	0.310000	234.000000	0.000000	374.000000	130.000000	1533.000000	222.000000	0.000000	52.000000	0.000000	329.000000	12.000000	0.000000	365.000000	363.000000	0.000000	0.000000	160.000000	0.000000	13.000000	326.000000	105.000000	0.000000	79.000000	255.000000	1.000000	4.000000	10.000000	5.000000	349.000000	2.000000	685.000000	1.000000	29.000000	442.000000	1.000000	45.000000	587.000000	0.000000	84.000000	0.000000	0.000000	1115.000000	110.000000	3.000000	197.000000	24.000000	25.000000	118.000000	0.000000	30.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
50%	444.000000	5.000000	53.000000	206.152870	0.000000	0.000000	0.000000	83547.999873	0.000000	0.000000	0.000000	3716.855123	0.000000	0.000000	848.220633	3925.995681	0.000000	0.000000	0.000000	1602.680584	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	3183.087137	2.210000	4.270000	7.640000	0.320000	0.320000	0.100000	0.200000	3.860000	0.058684	11.030000	0.200000	0.150000	0.400000	0.260000	0.450000	2.220000	0.830000	0.200000	0.060000	0.210000	7.370000	0.240000	0.050000	0.440000	0.320000	4.770000	0.210000	1.080000	0.140000	2.210000	0.300000	0.200000	0.150000	3.310000	0.080000	1.440000	0.210000	5.110000	1.480000	0.150000	0.100000	0.220000	0.220000	0.240000	3.840000	0.160000	0.440000	2.030000	0.450000	0.390000	0.200000	0.950000	2.980000	0.200000	0.280000	0.570000	0.420000	0.670000	0.470000	0.540000	1.530000	0.300000	0.420000	1.800000	0.070000	0.170000	2.720000	0.880000	0.470000	0.810000	0.090000	0.530000	0.190000	0.240000	1.370000	0.280000	0.140000	0.920000	0.690000	0.160000	0.090000	0.230000	0.490000	0.280000	9.110000	0.600000	0.143333	0.260000	0.260000	0.790000	3.360000	0.310000	4.970000	0.060000	2.950000	0.080000	0.220000	0.100000	1.060000	0.440000	2.070000	0.270000	0.090000	4.300000	2.280000	0.360000	0.140000	7.980000	0.270000	3.870000	1.540000	0.650000	0.080000	0.150000	1.200000	0.100000	0.340000	0.200000	0.165000	0.170000	0.690000	0.185000	0.900000	0.190000	0.220000	0.230000	2.180000	0.090000	2.710000	0.490000	0.190000	0.120000	0.275000	0.170000	0.26000	0.200000	0.150000	0.760000	4.360000	0.380000	0.170000	3.810000	0.510000	0.100000	0.540000	2.060000	0.230000	0.100000	0.210000	0.440000	0.570000	0.490000	3.790000	0.160000	1.030000	0.200000	0.870000	0.110000	0.410000	0.100000	0.150000	0.940000	0.170000	0.170000	0.310000	4.470000	0.510000	1.680000	4.140000	1.980000	0.180000	2.780000	0.200000	0.360000	0.380000	0.060000	0.090000	1.410000	0.090000	0.880000	0.105000	0.500000	0.820000	1.100000	0.630000	0.170000	0.110000	0.080000	1.320000	0.100000	1.130000	0.580000	0.196667	0.190000	0.840000	0.080000	6.810000	0.320000	1.020000	0.260000	0.280000	0.140000	0.170000	0.140000	0.710000	0.640000	0.710000	0.140000	0.320000	1.230000	1.570000	0.370000	0.220000	0.180000	1.100000	0.430000	1.460000	21.730000	5.470000	1.000000	0.700000	3.010000	0.380000	0.210000	0.290000	1.560000	0.430000	6.820000	0.480000	1.360000	1.310000	1.530000	0.143333	0.660000	0.066667	0.410000	1.040000	0.200000	0.480000	2.770000	0.230000	0.240000	3.960000	0.140000	0.160000	0.590000	0.250000	0.270000	0.430000	0.670000	0.330000	4.110000	0.230000	0.180000	0.990000	0.220000	0.910000	0.300000	0.730000	0.150000	0.260000	0.510000	0.223333	2.25000	0.070000	1.850000	0.090000	2.150000	0.140000	1.450000	0.200000	0.280000	0.220000	0.170000	0.270000	0.090000	0.14000	0.380000	0.470000	0.090000	5.120000	0.13000	0.490000	0.270000	0.200000	0.160000	0.820000	1.090000	0.390000	0.760000	34.680000	0.440000	1.47000	0.150000	0.160000	1.850000	0.560000	0.100000	0.690000	0.690000	0.810000	1.740000	0.140000	0.610000	0.050000	0.090000	0.166667	0.260000	0.170000	1.200000	0.890000	0.180000	0.060000	0.180000	0.155000	0.510000	0.200000	0.240000	0.580000	0.280000	0.250000	1.650000	0.100000	0.070000	0.360000	0.210000	0.590000	0.470000	0.410000	0.080000	0.140000	0.340000	0.160000	0.540000	0.080000	0.290000	0.390000	0.340000	0.165000	0.090000	2.480000	1.060000	0.280000	0.330000	0.230000	0.110000	0.590000	0.060000	0.770000	0.200000	0.780000	0.300000	1.980000	3.920000	6.050000	0.500000	1.630000	1.150000	0.340000	2.180000	0.830000	1.740000	0.090000	0.150000	0.170000	0.103200	0.340000	0.490000	0.070000	2.660000	0.200000	0.110000	0.370000	5.420000	0.560000	0.950000	0.670000	0.910000	0.090000	0.250000	0.150000	0.240000	0.230000	0.440000	0.163333	0.290000	0.850000	0.620000	0.970000	0.600000	0.140000	1.070000	3.960000	0.220000	0.580000	0.200000	0.200000	1.300000	4.130000	0.070000	0.820000	0.540000	2.790000	0.770000	0.470000	0.170000	0.070000	0.240000	1.040000	0.060000	0.210000	2.820000	1.220000	0.280000	0.420000	3.130000	0.270000	2.190000	0.430000	0.170000	0.470000	0.360000	718.000000	63.000000	976.000000	509.000000	3910.000000	600.000000	167.000000	160.000000	2470.000000	1295.000000	77.000000	286.000000	1709.000000	841.000000	378.000000	116.000000	641.000000	502.000000	43.000000	734.000000	672.000000	636.000000	699.000000	862.000000	112.000000	251.000000	405.000000	66.000000	1362.000000	185.000000	2926.000000	1363.000000	116.000000	1332.000000	438.000000	282.000000	1539.000000	85.000000	764.000000	92.000000	958.000000	4042.000000	444.000000	22.000000	937.000000	603.000000	219.000000	622.000000	47.000000	357.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
75%	1507.000000	22.000000	146.000000	321.141907	1324.223106	0.000000	0.000000	116581.162052	273.518026	0.000000	0.000000	6900.089940	0.000000	0.000000	1644.403817	6967.929272	0.000000	0.000000	0.000000	3907.900066	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	4770.463805	2.630000	4.540000	8.310000	0.370000	0.500000	0.140000	0.320000	4.140000	0.080000	12.620000	0.240000	0.170000	0.460000	0.300000	0.520000	2.600000	1.020000	0.236667	0.080000	0.360000	7.920000	0.310000	0.070000	0.530000	0.360000	5.170000	0.270000	1.200000	0.170000	2.680000	0.360000	0.250000	0.180000	4.060000	0.100000	1.610000	0.260000	5.480000	1.750000	0.170000	0.130000	0.280000	0.270000	0.280000	4.130000	0.180000	0.540000	2.210000	0.500000	0.440000	0.230000	1.080000	3.380000	0.240000	0.340000	0.670000	0.470000	0.770000	0.520000	0.740000	1.670000	0.340000	0.470000	2.040000	0.080000	0.200000	2.990000	0.950000	0.550000	0.890000	0.130000	0.630000	0.230000	0.300000	1.500000	0.350000	0.180000	1.060000	0.800000	0.193333	0.110000	0.270000	0.560000	0.330000	11.880000	0.680000	0.180000	0.300000	0.340000	0.960000	3.620000	0.360000	6.650000	0.080000	3.210000	0.100000	0.310000	0.120000	1.190000	0.500000	2.320000	0.310000	0.110000	4.670000	2.500000	0.430000	0.170000	9.230000	0.330000	4.170000	1.670000	0.740000	0.100000	0.190000	1.310000	0.130000	0.400000	0.240000	0.200000	0.200000	0.820000	0.220000	1.040000	0.236667	0.290000	0.280000	2.710000	0.110000	2.980000	0.550000	0.280000	0.140000	0.320000	0.200000	0.32000	0.260000	0.190000	0.890000	4.700000	0.450000	0.203333	4.520000	0.620000	0.130000	0.610000	2.210000	0.270000	0.140000	0.260000	0.490000	0.650000	0.550000	4.110000	0.195000	1.170000	0.233333	0.960000	0.140000	0.460000	0.130000	0.182500	1.050000	0.200000	0.220000	0.360000	4.870000	0.570000	1.980000	4.430000	2.330000	0.230000	3.000000	0.240000	0.440000	0.440000	0.090000	0.120000	1.520000	0.120000	1.020000	0.130000	0.560000	0.910000	1.250000	0.720000	0.220000	0.146667	0.110000	1.560000	0.120000	1.290000	0.670000	0.240000	0.230000	0.980000	0.110000	8.410000	0.400000	1.160000	0.316667	0.340000	0.160000	0.220000	0.165000	0.830000	0.730000	0.830000	0.170000	0.400000	1.380000	1.770000	0.510000	0.275000	0.240000	1.220000	0.510000	1.590000	24.700000	5.990000	1.110000	0.820000	3.270000	0.450000	0.260000	0.370000	1.770000	0.490000	8.440000	0.540000	1.700000	1.440000	1.700000	0.170000	0.730000	0.090000	0.480000	1.250000	0.240000	0.540000	2.990000	0.300000	0.290000	4.580000	0.170000	0.200000	0.650000	0.300000	0.321613	0.520000	0.760000	0.370000	4.420000	0.260000	0.230000	1.110000	0.260000	1.000000	0.360000	0.810000	0.180000	0.310000	0.570000	0.280000	2.49000	0.090000	2.270000	0.118889	2.350000	0.157500	1.560000	0.240000	0.330000	0.270000	0.200000	0.320000	0.120000	0.17000	0.470000	0.550000	0.120000	5.810000	0.15750	0.550000	0.320000	0.250000	0.190000	1.010000	1.310000	0.512000	0.930000	36.620000	0.490000	1.59000	0.190000	0.190000	2.000000	0.640000	0.130000	0.770000	0.790000	0.930000	1.990000	0.160000	0.780000	0.120000	0.114000	0.200000	0.330000	0.205000	1.610000	0.990000	0.220000	0.080000	0.225000	0.190000	0.580000	0.260000	0.325833	0.650000	0.360000	0.310000	1.840000	0.122500	0.100000	0.470000	0.260000	0.660000	0.540000	0.460000	0.098333	0.175000	0.390000	0.190000	0.600000	0.110000	0.350000	0.480000	0.400000	0.200000	0.120000	2.730000	1.170000	0.330000	0.380000	0.270000	0.130000	0.670000	0.080000	0.850000	0.230000	0.870000	0.350000	2.310000	4.290000	7.330000	0.580000	1.750000	1.350000	0.390000	2.370000	0.920000	2.220000	0.120000	0.170000	0.210000	0.130000	0.410000	0.585000	0.100000	3.080000	0.240000	0.140000	0.430000	5.880000	0.680000	1.050000	0.760000	1.050000	0.120000	0.300000	0.190000	0.280000	0.310000	0.550000	0.200000	0.330000	0.940000	0.730000	1.070000	0.660000	0.170000	1.300000	4.700000	0.260000	0.670000	0.290000	0.250000	1.460000	4.560000	0.090000	0.910000	0.610000	3.140000	0.870000	0.540000	0.225000	0.090000	0.310000	1.150000	0.094930	0.250000	3.070000	1.360000	0.340000	0.480000	3.470000	0.330000	2.520000	0.510000	0.200000	0.550000	0.420000	1705.000000	270.000000	2970.000000	1072.000000	8865.000000	1806.000000	786.000000	365.000000	7121.000000	3165.000000	156.000000	636.000000	3371.000000	2324.000000	960.000000	520.000000	1756.000000	1253.000000	266.000000	1158.000000	2037.000000	2576.000000	1698.000000	2074.000000	371.000000	663.000000	960.000000	363.000000	3421.000000	670.000000	6668.000000	2885.000000	344.000000	4262.000000	1237.000000	815.000000	4524.000000	371.000000	1669.000000	366.000000	2279.000000	8807.000000	1274.000000	131.000000	1844.000000	1584.000000	785.000000	2319.000000	200.000000	944.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
max	162871.000000	2441.000000	2580.000000	1602.563385	29828.486204	24400.214066	19336.781144	387777.167531	16652.730527	8285.291661	10025.855916	160126.276081	11887.411288	11793.697781	18856.420590	57234.221808	607.512334	10401.504845	3383.641439	64258.429881	12525.948910	11341.117042	659.513327	9458.974853	5443.848683	2072.314679	30513.293976	5.090000	7.560000	20.410000	0.950000	1.430000	1.090000	1.600000	12.820000	0.470000	35.300000	1.070000	0.450000	1.170000	0.820000	4.880000	3.490000	1.620000	0.890000	0.320000	3.470000	19.350000	0.970000	0.240000	1.390000	3.660000	27.960000	0.830000	10.100000	22.100000	5.730000	2.630000	1.320000	0.570000	5.630000	0.390000	4.930000	0.980000	8.440000	2.390000	0.450000	0.380000	3.670000	0.980000	0.950000	9.960000	0.500000	1.020000	3.470000	2.470000	2.460000	0.800000	1.610000	20.680000	1.590000	1.470000	1.460000	1.240000	1.300000	1.550000	1.950000	8.090000	0.980000	1.120000	6.370000	0.300000	0.600000	4.930000	1.900000	3.340000	1.660000	0.550000	1.400000	1.550000	1.030000	5.330000	2.250000	1.480000	3.550000	1.790000	3.310000	0.630000	1.160000	4.770000	3.030000	67.920000	4.060000	4.040000	1.150000	1.590000	1.990000	4.780000	2.300000	19.370000	0.290000	6.310000	0.380000	0.970000	0.710000	2.010000	1.440000	26.160000	1.070000	0.380000	13.170000	4.140000	0.970000	0.590000	21.390000	1.490000	14.330000	3.910000	1.320000	0.480000	0.740000	2.040000	0.410000	1.370000	0.890000	0.600000	0.500000	2.500000	1.120000	3.190000	3.060000	1.400000	3.430000	6.210000	0.520000	4.490000	1.520000	17.790000	0.450000	0.940000	0.730000	1.42000	3.530000	0.450000	1.760000	6.850000	1.300000	0.660000	28.480000	1.230000	0.390000	3.570000	7.070000	1.030000	0.390000	0.830000	1.140000	1.590000	1.050000	8.920000	0.460000	1.840000	0.790000	1.750000	0.400000	1.320000	0.400000	0.670000	2.430000	0.650000	0.700000	1.330000	8.590000	2.560000	10.260000	8.250000	6.570000	0.630000	9.030000	1.050000	1.970000	1.310000	2.310000	0.500000	2.530000	0.520000	5.710000	0.390000	1.240000	3.490000	2.020000	1.400000	0.820000	0.400000	0.940000	2.570000	1.180000	2.010000	1.180000	0.580000	1.040000	2.160000	0.470000	15.740000	2.070000	3.260000	1.030000	4.980000	1.550000	0.550000	0.520000	1.620000	2.940000	1.530000	0.430000	1.060000	2.260000	4.000000	2.790000	3.150000	1.350000	1.840000	1.270000	3.350000	100.000000	10.190000	4.670000	1.330000	5.540000	1.270000	1.170000	4.620000	2.610000	1.210000	15.630000	3.650000	3.200000	2.620000	3.540000	0.560000	1.740000	0.360000	1.390000	1.950000	0.590000	1.190000	4.940000	1.310000	3.580000	7.720000	1.150000	0.480000	2.310000	1.010000	0.880000	9.150000	5.570000	1.070000	8.290000	0.880000	13.110000	2.770000	0.890000	1.790000	1.140000	1.520000	0.590000	0.820000	4.760000	1.380000	5.15000	0.310000	5.120000	0.350000	3.400000	0.440000	2.730000	1.010000	0.830000	1.020000	0.510000	1.210000	0.430000	0.53000	1.180000	1.280000	0.440000	10.850000	0.48000	1.320000	0.900000	0.620000	0.610000	1.740000	2.230000	1.560000	1.640000	47.730000	1.230000	3.00000	0.410000	0.880000	3.520000	5.430000	0.430000	1.740000	1.650000	3.080000	5.410000	1.930000	2.940000	1.610000	0.360000	0.680000	0.900000	0.860000	5.170000	1.600000	1.980000	0.380000	2.500000	0.430000	2.250000	0.970000	0.900000	1.360000	1.120000	1.100000	3.940000	0.440000	0.380000	1.010000	0.900000	1.400000	1.190000	1.110000	0.390000	0.520000	1.210000	0.700000	1.550000	0.370000	0.920000	1.670000	2.090000	0.570000	0.400000	4.760000	1.970000	1.060000	1.050000	1.090000	0.520000	2.670000	0.310000	1.790000	2.060000	2.960000	1.010000	4.100000	7.820000	25.650000	1.530000	3.670000	2.470000	1.430000	4.300000	2.270000	5.710000	0.370000	0.550000	0.680000	0.450000	1.670000	1.280000	0.370000	7.750000	2.480000	1.410000	1.100000	8.700000	2.050000	2.120000	2.170000	2.030000	0.370000	0.890000	0.410000	1.160000	0.890000	1.450000	0.650000	3.510000	4.160000	1.940000	1.520000	1.630000	0.590000	4.750000	8.290000	1.770000	1.880000	2.390000	0.560000	2.450000	6.210000	0.340000	1.770000	1.400000	4.220000	1.450000	4.780000	1.230000	0.370000	0.910000	3.170000	0.400000	1.080000	14.870000	2.810000	0.990000	1.080000	4.740000	1.290000	10.530000	2.020000	0.620000	3.820000	1.450000	12972.000000	4013.000000	17234.000000	8434.000000	133669.000000	25224.000000	23678.000000	3729.000000	150251.000000	65039.000000	4789.000000	2527.000000	89195.000000	31431.000000	11885.000000	15972.000000	22912.000000	31161.000000	2149.000000	36646.000000	33090.000000	65557.000000	16184.000000	26182.000000	2216.000000	6135.000000	12443.000000	5082.000000	38461.000000	7437.000000	132093.000000	45901.000000	2807.000000	37626.000000	10073.000000	10433.000000	33650.000000	16228.000000	35951.000000	3047.000000	41464.000000	162871.000000	14754.000000	2779.000000	40246.000000	33069.000000	9164.000000	16956.000000	1541.000000	17525.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000

Data Science 2: Advanced Topics in Data Science

Final Project¶

Topic H: Dynamics of Disease Transmision and Human Behavior¶

Spring 2022¶

GROUP 10: Jaqueline Garcia-Yi, Rafael Hernandez, Hugo Munoz Sanchez, and Kerem Atalay¶

TABLE OF CONTENTS¶

I. Background and Introduction¶

II. Research question¶

III. Exploratory data analysis: trend graphs, correlations, and cluster analysis¶

IV. Model results and discussions¶

4.1. Preliminary: import commands, verify version of Tensorflow, and check GPU functioning¶

4.2. Read csv files¶

4.3. Merge files and clean data (handling of NAs)¶

4.4. Train-test split and normalization (per State) ¶

4.5. Time series baseline models¶

4.5.1 New York¶

a) Simple Moving Average Models¶

b) Auto Regressive Moving Average (ARMA) Model¶

4.5.2 California¶

a) Simple Moving Average Models¶

b) Auto Regressive Moving Average (ARMA) Model¶

4.6. RNN models¶

4.6.1 New York¶

a) One-layer RNN model¶

b) Multi-layer RNN model¶

4.6.2 California¶

a) One-layer RNN model¶

b) Multi-layer RNN model¶

4.7. LSTM models¶

4.7.1 New York¶

a) One-layer LSTM model¶

b) Multi-layer LSTM model¶

c) Two-layer LSTM model Using Time Steps¶

d) Multi-layer LSTM model Using Time Steps¶

4.7.2 California¶

a) One layer LSTM model¶

b) Multi-layer LSTM model¶

c) Two-layer LSTM model Using Time Steps¶

d) Multi-layer LSTM model Using Time Steps¶

V. Conclusions and way forward¶

VI. References¶

GROUP 10: Jaqueline Garcia-Yi, Rafael Hernandez, Hugo Munoz Sanchez, and Kerem Atalay

¶

I. Background and Introduction ¶

II. Research question ¶

III. Exploratory data analysis: trend graphs, correlations, and cluster analysis ¶

IV. Model results and discussions ¶

4.1. Preliminary: import commands, verify version of Tensorflow, and check GPU functioning ¶

4.2. Read csv files ¶

4.5. Time series baseline models ¶

4.5.1 New York ¶

4.5.2 California ¶

4.6. RNN models ¶

4.6.1 New York ¶

4.6.2 California ¶

4.7. LSTM models ¶

4.7.1 New York ¶

4.7.2 California ¶

V. Conclusions and way forward ¶

VI. References ¶