BlogChristmas.py

# -*- coding: utf-8 -*-
"""
Analyze Christmas spurt.
@author: Matt Bayer
"""



import os
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import plotly.express as px
import datetime

import plotly.subplots
import plotly.express as px
import plotly.graph_objects as go
from plotly.offline import plot as pplot
plotpath = '.\\docs\\assets\\plotly'

import covid

#%% Load data

state = covid.read_covid_data_wi('state')

# rename
col_rename = {'Date': 'Date', 'POS_NEW': 'Cases', 'TEST_NEW': 'Tests', 'DTH_NEW': 'Deaths', 'HOSP_NEW': 'Hospitalizations'}
state = state.rename(columns=col_rename)

# tests by result date
tests = covid.read_bytest_wi("data\\By_Test_Data_data_2021-01-06.csv")


#%% Process tests by result date

tests['Tests 7-day'] = tests['Tests'].rolling(7).mean()
tests['Positives 7-day'] = tests['Positives'].rolling(7).mean()
tests['Positivity 7-day'] = tests['Positives 7-day'] / tests['Tests 7-day']
tests['Positivity'] = tests['Positives'] / tests['Tests']
tests['Positivity 7-day x20000'] = tests['Positivity 7-day'] * 20000
tests['Prevalence Index'] = np.sqrt(tests['Positives 7-day'] * tests['Positivity 7-day']) * 200

tests['Weekday'] = tests.Date.apply(lambda d: d.weekday())

tests.plot(x='Date', y=['Positives 7-day', 'Positivity 7-day x20000', 'Prevalence Index'])

# Bar plot of tests and positives, like DHS
# fig = px.bar(tests, x='Date', y=['Positives', 'Tests'], barmode='group')
# pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\temp1.html')

# Make separate lines for each day of the week
# fig = px.line(tests, x='Date', y='Positives', color='Weekday')
# pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\temp2.html')



#%% Plot tests by result date

# remember below shows 7-day average positivity as average of the positivity, 
# not more correctly positivity of the averages.

covid.plotly_twolines(
    tests, 
    'Positivity', 
    'Positives', 
    plotcolors=['violet', 'steelblue', 'thistle'],
    secondary_scale=3e4,
    savefile=plotpath+'\\Pos-Positivity-ByResult-WI.html',
    )

# covid.plotly_casetest(sourcedata=tests, 
#                       case_col='Positives', 
#                       test_col='Tests', 
#                       date_col='Date', 
#                       savefile=plotpath + '\\Pos-Test-ByResult-WI.html',
#                       )


#%% Cases by test date for Wisconsin
filename = 'C:\dev\covid-wisconsin\data\Cases_with_prob_stacked_data_2021-01-08.csv'

wi = pd.read_csv(filename)
# filter out redundant data
wi = wi.loc[wi['Measure Names'] == 'Confirmed cases']  
# rename columns
col_rename = {'Day of Epi Dt': 'Date', 'Stacked Confirm + Probable cases': 'Cases'}
wi = wi[col_rename.keys()]
wi = wi.rename(columns=col_rename)
wi['Date'] = pd.to_datetime(wi['Date'])

#%% cases bar plot

wi = wi.loc[wi['Date'] >= datetime.datetime(2020,12,7)]
wi = wi.loc[wi['Date'] <= datetime.datetime(2021,1,3)]

fig = px.bar(
    wi, 
    x='Date', 
    y='Cases', 
    color_discrete_sequence=['steelblue'],    
    title='Cases by Symptom/Test Date - Wisconsin',
    width=700,
    height=340,
    )

fig.update_layout(showlegend=False)

# add dividers
date = datetime.datetime(2020,12,13,12)
delta = datetime.timedelta(days=7)
dividers = list()
for d in range(0,3):
    dividers.append(
        dict(
            type= 'line', line_color='gray', line_dash='dot',
            yref= 'paper', y0= 0, y1= 1,
            xref= 'x', x0=date, x1=date
        )
    )
    date = date + delta

fig.update_layout(shapes=dividers)
fig.add_annotation(text='Christmas',
                   x=datetime.datetime(2020,12,25,0), 
                   y=1400, 
                   yanchor='bottom', xanchor='center', showarrow=False, textangle=270)
                   
pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\Christmas-WI.html')

save_png = '.\\docs\\assets\\Christmas-WI.png'
fig.write_image(
    save_png,
    width=700,
    height=340,
    engine='kaleido',
)
os.startfile(save_png)

#%% Christmas surge for Milwaukee

# get milwaukee test numbers by copying html of the dashboard and scraping it
# get this originally by doing a browser html inspection of the graph on the Milwaukee county dashboard,
# then copying it to a text file.  The tooltips of the graph are stored in "aria-label" labels, so
# I can regex on these labels below.


html_cases = 'C:\dev\covid-wisconsin\data\Dashboard-Milwaukee-Cases_2021-01-08.html'
html_tests = 'C:\dev\covid-wisconsin\data\Dashboard-Milwaukee-Tests_2021-01-08.html'

mke_case = covid.read_dashboard_mke(html_cases)
mke_test = covid.read_dashboard_mke(html_tests)
mke_test = mke_test.replace(to_replace='7 Day Average', value='Tests 7-day')

mke = mke_case.append(mke_test)
mke = mke.pivot(index='Date', columns='variable', values='value')
mke = mke.reset_index()

mke['Weekday'] = mke['Date'].apply(lambda d: d.weekday())
mke['Positivity'] = mke['Cases'] / mke['Tests']

# sum weekly Mon-Sun
# sum with a rolling weekly window, then keep the results from Sunday
weekly = mke[['Cases', 'Tests']].rolling(7).sum()
weekly['Date'] = mke['Date']
weekly = weekly.loc[mke['Weekday'] == 6]

# sum weekly Mon-Wed only
# sum with a rolling 3-day window, then keep the results from Wednesday
monwed = mke[['Cases', 'Tests']].rolling(3).sum()
monwed['Date'] = mke['Date']
monwed = monwed.loc[mke['Weekday'] == 2]
monwed['Positivity'] = monwed['Cases'] / monwed['Tests'] * 100 # in %


weekly['Positivity'] = weekly['Cases'] / weekly['Tests']

weekly.plot(x='Date', y='Positivity', marker='.')


#%% facet bar plot?

temp = mke.loc[mke['Date'] >= datetime.datetime(2020,12,7)]
temp = temp.loc[temp['Date'] <= datetime.datetime(2021,1,3)]
temp = temp[['Date', 'Cases', 'Tests']].melt(id_vars='Date')

fig = px.bar(
    temp, 
    x='Date', 
    y='value', 
    facet_row='variable',
    color='variable',
    color_discrete_sequence=['steelblue', 'olivedrab'],    
    title='Cases and Tests by Test Date - Milwaukee',
    width=700,
    height=500,
    )

fig.for_each_annotation(lambda a: a.update(text=''))
fig.update_layout(showlegend=False)
fig.update_yaxes(matches=None)
fig.update_yaxes(title='Cases')
fig.update_yaxes(title='Tests', row=1)


# add dividers
date = datetime.datetime(2020,12,13,12)
delta = datetime.timedelta(days=7)
dividers = list()
for d in range(0,3):
    dividers.append(
        dict(
            type= 'line', line_color='gray', line_dash='dot',
            yref= 'paper', y0= 0, y1= 1,
            xref= 'x', x0=date, x1=date
        )
    )
    date = date + delta

fig.update_layout(shapes=dividers)
fig.add_annotation(text='Christmas',
                   x=datetime.datetime(2020,12,24,23), 
                   y=700, 
                   yanchor='bottom', xanchor='center', showarrow=False, textangle=270)
                   
fig.add_annotation(text='Christmas',
                   x=datetime.datetime(2020,12,24,23), 
                   y=50, row=0, col=0,
                   yanchor='bottom', xanchor='center', showarrow=False, textangle=270)

pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\Christmas-Milwaukee.html')

save_png = '.\\docs\\assets\\Christmas-Milwaukee.png'
fig.write_image(
    save_png,
    width=700,
    height=500,
    engine='kaleido',
)
os.startfile(save_png)


#%% Mon-Wed line plot
# fig = px.bar(weekly, x='Date', y=['Cases', 'Tests'], barmode='group')
# pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\temp.html')

monwed['Week of'] = monwed['Date'].apply(lambda d: d.strftime('%b %d'))
plotdata = monwed.iloc[-5:].reset_index(drop=True)
plotdata.loc[3,'Week of'] = 'Christmas'



fig = plotly.subplots.make_subplots(
    specs=[[{'secondary_y': True}]]
    )

fig.add_trace(
    go.Scatter(
        x=plotdata['Week of'], 
        y=plotdata.Cases, 
        name='Cases', 
        line_color='steelblue', 
        ),
    secondary_y=False,
    )

# add positivity on secondary axis
fig.add_trace(
    go.Scatter(
        x=plotdata['Week of'], 
        y=plotdata.Positivity, 
        name='Positivity', 
        line_color='violet', 
        ),
    secondary_y=True,
    )
    

fig.update_layout(title='Milwaukee Mon-Wed Sums')

fig.update_traces(mode='lines+markers', marker_size=10, line_dash='dot')
fig.update_xaxes(title='Week of...')
fig.update_yaxes({'range': [0, 4500]}, title='Cases', secondary_y=False)
fig.update_yaxes({'range': [0, 20]}, title='Positivity (%)', secondary_y=True)


pplot(fig, include_plotlyjs='cdn', filename=plotpath+'\\Christmas-MonWed-Milwaukee.html')

save_png = '.\\docs\\assets\\Christmas-MonWed-Milwaukee.png'
fig.write_image(
    save_png,
    width=700,
    height=400,
    engine='kaleido',
)
os.startfile(save_png)