Analysis of COVID-19 pandemic in INDIA¶
What is Corona Virus(COVID-19)?¶
Coronavirus is a family of viruses that can cause illness varying from common cold and cough to sometimes more severe disease. SARS- CoV-2 (n-coronavirus) is the new virus of the coronavirus family, which was first discovered in 2019 and has not been identified in humans before. It is a contiguous virus which started from Wuhan in December 2019 and later, it was declared as a Pandemic by WHO due to high rate spreads throughout the world. As the Pandemic is spreading all over the world; it is important to understand about this spread. This Notebook is an effort to analyze the cumulative data of confirmed cases, deaths, and recovered cases over time. In this notebook, the focus is to analyze the trend of this virus all over India.
History of COVID-19 in India¶
India reported its first case of COVID-19 in Kerala on January 30, which rose to three cases by February 3; These were students who had returned from Wuhan, China. No significant rise in cases was seen in the rest of February. India observed a 14-hour voluntary public curfew at the instance of the prime minister Narendra Modi on 22 March 2020. The government followed it up with lockdowns in 75 districts where COVID cases had occurred as well as all major cities. Further, on 24 th March, the prime minister ordered a nationwide lockdown for 21 days that affected the entire 1.3 billion population of India. The transmission increased rapidly during March, after several cases were reported all over the country, most of which were linked to people with a travel history to affected countries. On 12 th March, a 76-year-old man who had returned from Saudi Arabia became the first victim of the virus in the country. On 4 th March, 22 new cases came to light, including those of an Italian tourist group with 14 infected members. But number of cases increased dramatically after 19 th March. Experts suggest the number of infections might be much higher as India's testing rates are among the lowest in the world. The infection rate of COVID-19 in India is reported to be 1.7, significantly lower than in the worst affected countries.
from IPython.display import *
Image("https://media.foxbusiness.com/BrightCove/854081161001/202003/2652/854081161001_6141155653001_6141149610001-vs.jpg")
Data collection and preprocessing¶
import warnings
warnings.filterwarnings("ignore")
!pip install --upgrade plotly
import json
token = {"username":"abshkpskr","key":"7bf37d08aafdc81ecb2640a1960fddbc"}
with open('kaggle.json', 'w') as file:
json.dump(token, file)
!mkdir ~/.kaggle
!cp kaggle.json ~/.kaggle/
!chmod 600 ~/.kaggle/kaggle.json
!kaggle datasets download -d imdevskp/covid19-corona-virus-india-dataset
from zipfile import ZipFile
ZipFile('covid19-corona-virus-india-dataset.zip').extractall()
import os
for i in sorted(os.listdir()):
print(i)
#libraries for data management
import numpy as np
import pandas as pd
#libraries for visualization
import matplotlib.pyplot as plt
import seaborn as sns
import plotly
import plotly.express as px
import plotly.graph_objs as go
import folium
from google.colab import files
from IPython.display import *
# plotly.io.renderers.default = 'colab'
plotly.offline.init_notebook_mode(connected=True)
data = pd.read_csv('complete.csv')
district_data = pd.read_csv('district_level_latest.csv')
state_data = pd.read_csv('state_level_latest.csv')
nation_data = pd.read_csv('nation_level_daily.csv')
patients_data = pd.read_csv('patients_data.csv')
tests_data = pd.read_csv('tests_day_wise.csv')
state_tests_data = pd.read_csv('tests_state_wise.csv')
zones = pd.read_csv('https://raw.githubusercontent.com/AbshkPskr/Covid-19-Analysis/master/India%20Covid%20Zones.csv')
data
data = data.drop(['New cases','New deaths','New recovered'],axis = 1)
data.columns = ['date','state','lat','long','confirmed','deaths','recovered']
data = data[['lat','long','state','date','confirmed','deaths','recovered']]
data.loc[(data['state'] == 'Union Territory of Ladakh'),'state'] = 'Ladakh'
data.loc[(data['state'] == 'Union Territory of Jammu and Kashmir'),'state'] = 'Jammu and Kashmir'
data.loc[(data['state'] == 'Telengana'),'state'] = 'Telangana'
data.loc[(data['state'] == 'Odisha'),'state'] = 'Orissa'
data.loc[(data['state'] == 'Union Territory of Chandigarh'),'state'] = 'Chandigarh'
bing_data = pd.read_csv('https://raw.githubusercontent.com/microsoft/Bing-COVID-19-Data/master/data/Bing-COVID19-Data.csv',index_col = 'ID')
bing_data = bing_data[(bing_data['Country_Region'] == 'India')]# & (bing_data['Latitude'] != 22.49312) & (bing_data['Longitude'] != 79.72701)]
bing_data = bing_data.drop(['ISO2','ISO3','Country_Region'],axis = 1)
bing_data.columns = ['date','confirmed','new_confirmed','deaths','new_deaths','recovered','new_recovered','lat','long','state','city']
bing_data['active'] = bing_data['confirmed'] - (bing_data['deaths'] + bing_data['recovered'])
bing_data = bing_data[(bing_data['state'] == 'Chandigarh') |
(bing_data['state'] == 'Andaman and Nicobar Islands') |
(bing_data['city'].notna())]
# bing_data = bing_data[(bing_data['city'].isna())]
bing_data['city'] = bing_data['city'].replace(np.nan,"(Union Territory)")
bing_data = bing_data.replace(np.nan,0)
case_data = pd.DataFrame()
states = data['state'].unique()
for state in states:
temp = data[(data['state'] == state)]
temp1 = pd.DataFrame()
temp1 = temp1.append(temp.iloc[0:1,:],ignore_index=True)
temp1['new_confirmed'] = temp1['confirmed']
temp1['new_deaths'] = temp1['deaths']
temp1['new_recovered'] = temp1['recovered']
for i in range(1,len(temp)):
lat,lon,state,date,confirmed,death,recovered = temp.iloc[i].values
new_confirmed = temp.iloc[i,4] - temp.iloc[i-1,4]
new_deaths = temp.iloc[i,5] - temp.iloc[i-1,5]
new_recovered = temp.iloc[i,6] - temp.iloc[i-1,6]
if new_confirmed < 0:
confirmed = temp.iloc[i-1][4]
new_confirmed = 0
if new_deaths < 0:
deaths = temp.iloc[i-1][5]
new_deaths = 0
if new_recovered < 0:
recovered = temp.iloc[i-1][6]
new_recovered = 0
temp1.loc[i] = [lat,lon,state,date,confirmed,death,recovered,new_confirmed,new_deaths,new_recovered]
case_data = pd.concat([case_data,temp1],ignore_index=True)
#case_data
# case_data = bing_data[['date','state','city','confirmed','new_confirmed','deaths','new_deaths','recovered','new_recovered','lat','long']]
case_data['active'] = case_data['confirmed'] - (case_data['deaths'] + case_data['recovered'])
case_data['date'] = pd.to_datetime(case_data['date'])
case_data['confirmed'] = case_data['confirmed'].astype('int')
case_data['deaths'] = case_data['deaths'].astype('int')
case_data['recovered'] = case_data['recovered'].astype('int')
case_data['active'] = case_data['active'].astype('int')
case_data['new_confirmed'] = case_data['new_confirmed'].astype('int')
case_data['new_deaths'] = case_data['new_deaths'].astype('int')
case_data['new_recovered'] = case_data['new_recovered'].astype('int')
# case_data.loc[(case_data['state'] == ' Azerbaijan'),'state'] = 'Azerbaijan'
Data analysis¶
The following analysis techniques are used to depict how the coronavirus pandemic is affecting various demographic groups in India.
General Analysis¶
state_case_data = case_data.groupby(['state','date']).sum().reset_index()
state_case_data = state_case_data[state_case_data['date'] == max(state_case_data['date'].values)]
state_case_data = state_case_data.drop('date',axis = 1)
state_case_data = state_case_data.set_index('state')
state_case_data = state_case_data[['confirmed','new_confirmed','deaths','new_deaths',
'recovered','new_recovered','active']]
state_case_data['mortality rate'] = (state_case_data['deaths']/state_case_data['confirmed'])*100
state_case_data = state_case_data.rename(columns={'confirmed':'| confirmed |','new_confirmed':'| New confirmed |',
'deaths':'| deaths |','new_deaths':'| New deaths |',
'recovered':'| recovered |','new_recovered':'| New recovered |',
'active':'| Active |','mortality rate':'| Mortality Rate |'})
state_case_data.sort_values('| confirmed |', ascending= False).style\
.background_gradient(cmap='binary',subset=["| confirmed |"])\
.background_gradient(cmap='Blues',subset=["| New confirmed |"])\
.background_gradient(cmap='binary',subset=["| deaths |"])\
.background_gradient(cmap='Reds',subset=["| New deaths |"])\
.background_gradient(cmap='binary',subset=["| recovered |"])\
.background_gradient(cmap='Greens',subset=["| New recovered |"])\
.background_gradient(cmap='Purples',subset=["| Active |"])\
.background_gradient(cmap='YlOrBr',subset=["| Mortality Rate |"])
national_cumulative = pd.DataFrame(case_data.groupby(['date']).sum()).reset_index()
national_cumulative = national_cumulative[national_cumulative['date'] > '2020-03-01']
fig = go.Figure()
fig.add_trace(go.Scatter(x=national_cumulative['date'],y=national_cumulative['confirmed'],mode='lines',name='confirmed',line=dict( width=4)))
fig.add_trace(go.Scatter(x=national_cumulative['date'],y=national_cumulative['deaths'],mode='lines',name='deaths',line=dict( width=4)))
fig.add_trace(go.Scatter(x=national_cumulative['date'],y=national_cumulative['recovered'],mode='lines',name='recovered',line=dict( width=4)))
fig.add_trace(go.Scatter(x=national_cumulative['date'],y=national_cumulative['active'],mode='lines',name='Active',line=dict( width=4)))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.01,y=.98),
title_text = 'Number of COVID-19 cases nationwide',font_size=15,
xaxis_title="Date",
yaxis_title="Number of cases",
shapes=[dict(type="rect",xref="x",yref="paper",x0="2020-03-25",y0=0,x1="2020-04-14",y1=1,fillcolor="#D89393",opacity=0.5,layer="below",line_width=0,),
dict(type="rect",xref="x",yref="paper",x0="2020-04-14",y0=0,x1="2020-05-03",y1=1,fillcolor="#DA7676",opacity=0.5,layer="below",line_width=0,),
dict(type="rect",xref="x",yref="paper",x0="2020-05-03",y0=0,x1="2020-05-17",y1=1,fillcolor="#DD4747",opacity=0.5,layer="below",line_width=0,),
dict(type="rect",xref="x",yref="paper",x0="2020-05-17",y0=0,x1="2020-05-31",y1=1,fillcolor="#DF1818",opacity=0.5,layer="below",line_width=0,),
dict(type="rect",xref="x",yref="paper",x0="2020-05-31",y0=0,x1=max(national_cumulative['date']),y1=1,fillcolor="#A0DE8A",opacity=0.5,layer="below",line_width=0,)],
annotations=[dict(x="2020-04-05",y=max(national_cumulative['confirmed'])-10000,xref="x",yref="y",text="First Lockdown",showarrow=False),
dict(x="2020-04-23",y=max(national_cumulative['confirmed'])-10000,xref="x",yref="y",text="Second Lockdown",showarrow=False),
dict(x="2020-05-10",y=max(national_cumulative['confirmed'])-10000,xref="x",yref="y",text="Third Lockdown",showarrow=False),
dict(x="2020-05-24",y=max(national_cumulative['confirmed'])-10000,xref="x",yref="y",text="Fourth Lockdown",showarrow=False),
dict(x="2020-05-24",y=max(national_cumulative['confirmed'])-10000,xref="x",yref="y",text="Fourth Lockdown",showarrow=False),
])
fig.layout.hovermode = 'x'
fig.show()
national_daily_cumulative = pd.DataFrame(case_data.groupby(['date']).sum())
fig = go.Figure()
fig.add_trace(go.Scatter(x=national_daily_cumulative.index,y=national_daily_cumulative['new_confirmed'],mode='lines',name='confirmed',line=dict( width=4)))
fig.add_trace(go.Scatter(x=national_daily_cumulative.index,y=national_daily_cumulative['new_deaths'],mode='lines',name='deaths',line=dict( width=4)))
fig.add_trace(go.Scatter(x=national_daily_cumulative.index,y=national_daily_cumulative['new_recovered'],mode='lines',name='recovered',line=dict( width=4)))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.01,y=.98),
title_text = 'Number of daily COVID-19 cases nationwide',font_size=15,
xaxis_title="Date",
yaxis_title="Number of new cases",)
fig.layout.hovermode = 'x'
fig.show()
Map Visualization¶
def CreateMap(data,color,fill_color,location = [22,80],zoom = 5,dot_size = 2):
_map = folium.Map(location=location, tiles="OpenStreetMap", zoom_start=zoom)
for name,cases,deaths,lat,lon in data.values:
folium.CircleMarker([lat,lon],radius=((int(np.log(cases + 10)))*dot_size),color=color,fill_color=fill_color,
tooltip = "<h5 style='text-align:center;font-weight: bold'>"+ name +"</h5>"+
"<hr style='margin:10px;'>"+
"<ul style='color: #444;list-style-type:circle;align-item:left;padding-left:20px;padding-right:20px'>"+
"<li>confirmed: "+str(cases)+"</li>"+
"<li>deaths: "+str(deaths)+"</li>"+
"</ul>",fill_opacity=0.7).add_to(_map)
return _map
state_map_plot = case_data[case_data['date'] == max(case_data['date'].values)]
state_map_plot = state_map_plot[['state','confirmed','deaths','lat','long']].reset_index().drop('index',axis = 1)
CreateMap(state_map_plot,'#1C6179','#2C94B8',)
district_map_plot = bing_data.groupby(['state','city']).max().reset_index()[['state','city','confirmed','deaths','lat','long']]
# district_map_plot = bing_data[bing_data['date'] == max(bing_data['date'].values)]
district_map_plot['location'] = district_map_plot['city'] + ", " + district_map_plot['state']
district_map_plot = district_map_plot[['location','confirmed','deaths','lat','long']]
CreateMap(district_map_plot,'#1C6179','#2C94B8',dot_size = 1)
CreateMap(district_map_plot,'#1C6179','#2C94B8',[28.5,77.3],10,4)
HTML("<img src='https://upload.wikimedia.org/wikipedia/commons/thumb/9/95/COVID-19_India_Total_Cases_Animated_Map.gif/527px-COVID-19_India_Total_Cases_Animated_Map.gif'>")
# timeline_plot = case_data.groupby(['state','date']).sum().reset_index().sort_values(['date'])
# timeline_plot['date'] = timeline_plot['date'].astype('str')
# fig = px.choropleth(timeline_plot, locations="state",
# color=np.log10(timeline_plot["confirmed"]),
# hover_name="state",
# hover_data=["confirmed",'deaths','recovered'],
# color_continuous_scale=px.colors.sequential.thermal_r,
# locationmode="IN",
# animation_frame='date',color_continuous_midpoint = 3)
# fig.update_layout(margin=dict(l=20,r=0,b=0,t=70,pad=0),
# paper_bgcolor="white",
# height= 700,
# title_text = 'Number of daily COVID-19 cases worldwide',font_size=18)
# fig.show()
Race Map Visualization¶
case_confirmed_racemap = case_data[['state','date','confirmed']]
case_confirmed_racemap= case_confirmed_racemap.pivot_table(index='state',columns='date',values= 'confirmed')
case_confirmed_racemap.to_csv('country.csv')
files.download('country.csv')
%%HTML
<div class="flourish-embed flourish-bar-chart-race" data-src="visualisation/2554556" data-url="https://flo.uri.sh/visualisation/2554556/embed"><script src="https://public.flourish.studio/resources/embed.js"></script></div>
Comparison between most affected states¶
case_country = case_data.groupby(['state','date']).sum().reset_index()
countries = case_country[case_country['confirmed'] > 5000]['state'].unique()
def CreateComparisonPlot(attr,title,x_title,y_title, y_axis_type = None):
fig = go.Figure()
for i in countries:
if y_axis_type == 'log' :
one_country = case_country[(case_country['state'] == i) & (case_country[attr] > 0)][['date',attr]]
one_country['date'] = [i for i in range(1,len(one_country)+1)]
else:
one_country = case_country[(case_country['state'] == i)][['date',attr]]
fig.add_trace(go.Scatter(x=one_country['date'],
y=one_country[attr],mode='lines',
name=i,line=dict( width=4)))
# fig.add_annotation(text="First case of United states",
# x='2020-01-23', y=1, arrowhead=0, showarrow=True)
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.01,y=.98,font=dict(size =12)),
title_text = title,font_size=15,
xaxis_title=x_title,
yaxis_title=y_title)
if y_axis_type == 'log' :
fig.update_yaxes(type="log")
fig.update_layout(legend=dict(x=.85,y=.02,font=dict(size =12)))
fig.layout.hovermode = 'x'
fig.show()
CreateComparisonPlot('confirmed','Number of confirmed cases for most affected states','Date','Number of cases')
CreateComparisonPlot('deaths','Number of deaths in most affected states','Date','Number of cases')
CreateComparisonPlot('recovered','Number of recovered cases in most affected states','Date','Number of cases')
CreateComparisonPlot('new_confirmed','Number of Daily confirmed cases in most affected states','Date','Number of cases')
CreateComparisonPlot('new_deaths','Number of Daily Death cases in most affected states','Date','Number of cases')
CreateComparisonPlot('confirmed','Number of confirmed cases for most affected countries',
'Day','Number of cases (log scale)','log')
CreateComparisonPlot('deaths','Number of confirmed cases for most affected countries',
'Day','Number of cases (log scale)','log')
CreateComparisonPlot('recovered','Number of confirmed cases for most affected countries',
'Day','Number of cases (log scale)','log')
Trend of each state separately¶
case_state= case_data.groupby(['state','date']).sum().reset_index()
case_state['date'] = case_state['date'].astype('str')
countries = case_state[case_state['confirmed'] > 200].sort_values(['confirmed'],ascending=False)['state'].unique()
# row = int((len(countries)+1)/3)
rows = len(countries)
columns = 2
f = plt.figure(figsize=(20,rows*5))
gs = f.add_gridspec(rows,columns)
sns.set(style = "whitegrid")
country = 0
for i in range(0,rows):
data = case_state[case_state['state'] == countries[country]]
data = data[data['confirmed'] > 0]
data['day'] = [i for i in range(1,len(data)+1)]
for j in range(0,columns):
#if country == len(countries): break
ax = f.add_subplot(gs[i,j])
if j == 0:
sns.scatterplot(data = data,x = 'day',y = 'confirmed',s = 50,color="#4348C4",edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'confirmed', color = '#4348C4')
plt.fill_between(data['day'], data['confirmed'], alpha=0.30, color = '#4348C4')
sns.scatterplot(data = data,x = 'day',y = 'recovered',s= 50,color="#5BAC4D",edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'recovered',color="#5BAC4D")
plt.fill_between(data['day'], data['recovered'],alpha=0.30,color="#5BAC4D")
sns.scatterplot(data = data,x = 'day',y = 'deaths',s= 50,color='#BB3535',edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'deaths',color='#BB3535')
plt.fill_between(data['day'], data['deaths'],alpha=0.30,color='#BB3535')
ax.legend(['confirmed','recovered','deaths'], fontsize=16,loc = 'upper left')
ax.set_xlabel('Day',fontsize=20)
ax.set_ylabel('No. of cases',fontsize=20)
ax.set_title(countries[country], fontdict={'fontsize': 25, 'weight' : 'bold'}, color="black",loc = 'left')
text = str(data['date'].values[0]) + " - " + str(data['date'].values[-1]) + "\n" + "\n"
text += 'Total Cases :-' + "\n"
text += 'confirmed :' + str(data['confirmed'].values[-1]) + "\n"
text += 'deaths :' + str(data['deaths'].values[-1]) + "\n"
text += 'recovered :' + str(data['recovered'].values[-1]) + "\n" + "\n"
ax.text(0.02, 0.15, text, fontsize=15, transform=ax.transAxes,bbox=dict(facecolor='#F8F8F8', alpha=.7))
if j == 1:
sns.scatterplot(data = data,x = 'day',y = 'new_confirmed',s = 50,color="#4348C4",edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'new_confirmed', color = '#4348C4')
plt.fill_between(data['day'], data['new_confirmed'], alpha=0.30, color = '#4348C4')
sns.scatterplot(data = data,x = 'day',y = 'new_recovered',s= 50,color="#5BAC4D",edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'new_recovered',color="#5BAC4D")
plt.fill_between(data['day'], data['new_recovered'],alpha=0.30,color="#5BAC4D")
sns.scatterplot(data = data,x = 'day',y = 'new_deaths',s= 50,color='#BB3535',edgecolor = 'none')
sns.lineplot(data = data,x = 'day',y = 'new_deaths',color='#BB3535')
plt.fill_between(data['day'], data['new_deaths'],alpha=0.30,color='#BB3535')
ax.legend(['Daily confirmed','Daily recovered','Daily deaths'], fontsize=16,loc = 'upper left')
ax.set_xlabel('Day',fontsize=20)
ax.set_ylabel('No. of cases',fontsize=20)
text = 'In last 24 hrs :-' + "\n"
text += 'confirmed :' + str(data['new_confirmed'].values[-1]) + "\n"
text += 'deaths :' + str(data['new_deaths'].values[-1]) + "\n"
text += 'recovered :' + str(data['new_recovered'].values[-1])
ax.text(0.02, 0.3, text, fontsize=15, transform=ax.transAxes,bbox=dict(facecolor='#F8F8F8', alpha=.7))
country += 1
f.tight_layout()
Age Analysis¶
def GetAgeData(data,feature_name):
for i in data.index:
age = data.loc[data.index == i,feature_name]
try:
age = int(age)
except:
continue
if age <= 10: data.loc[data.index == i,feature_name] = '1-10'
if 11 <= age <= 20: data.loc[data.index == i,feature_name] = '11-20'
if 21 <= age <= 30: data.loc[data.index == i,feature_name] = '21-30'
if 31 <= age <= 40: data.loc[data.index == i,feature_name] = '31-40'
if 41 <= age <= 50: data.loc[data.index == i,feature_name] = '41-50'
if 51 <= age <= 60: data.loc[data.index == i,feature_name] = '51-60'
if 61 <= age <= 70: data.loc[data.index == i,feature_name] = '61-70'
if age > 70: data.loc[data.index == i,feature_name] = '70+'
return data.groupby(feature_name).size().reset_index(name = 'freq')
# patients_data.loc[(patients_data['Age Bracket'] == '1.5'),'Age Bracket'] = np.nan
# patients_data.loc[(patients_data['Age Bracket'] == '0.4'),'Age Bracket'] = np.nan
# patients_data.loc[(patients_data['Age Bracket'] == '0.3'),'Age Bracket'] = np.nan
patients_data.loc[(patients_data['Age Bracket'] == '28-35'),'Age Bracket'] = 31
patients_data.loc[(patients_data['Age Bracket'] == '6 Months'),'Age Bracket'] = 0.5
patients_data['Age Bracket'] = patients_data['Age Bracket'].astype('float')
age_data = GetAgeData(pd.DataFrame(patients_data['Age Bracket'].values,columns= ['Age Bracket']),'Age Bracket')
fig = go.Figure().add_trace(go.Pie(values=age_data['freq'],labels= age_data['Age Bracket'],hole = .5,text=age_data['Age Bracket']))
fig.update_layout(margin=dict(l=0,r=20,b=20,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'Percentage of cases recorded by age',font_size=20,
xaxis_title="Country",
yaxis_title="Rate")
fig.layout.hovermode = 'x'
fig.show()
age_data_deseased = GetAgeData(pd.DataFrame(patients_data[patients_data['Current Status'] == 'Deceased']['Age Bracket'].values,columns= ['Age Bracket']),'Age Bracket')
fig = go.Figure().add_trace(go.Pie(values=age_data_deseased['freq'],labels= age_data_deseased['Age Bracket'],
hole = .5,text=age_data_deseased['Age Bracket']))
fig.update_layout(margin=dict(l=0,r=20,b=20,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'Percentage of cases resulted in death by age',font_size=20,
xaxis_title="Country",
yaxis_title="Rate")
fig.layout.hovermode = 'x'
fig.show()
Gender Analysis¶
temp = pd.DataFrame(patients_data['Gender'].values,columns= ['Gender'])
temp = temp.fillna('NA')
temp.loc[(temp['Gender'] == 'Non-Binary'),'Gender'] = np.nan
gender_data = temp.groupby('Gender').size().reset_index(name = 'size')
fig = go.Figure().add_trace(go.Pie(values = gender_data['size'],labels = gender_data['Gender'],
hole = .5,text=gender_data['Gender']))
fig.update_layout(margin=dict(l=0,r=20,b=20,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'Percentage of cases recorded by gender',font_size=20,
xaxis_title="Country",
yaxis_title="Rate")
fig.show()
temp = pd.DataFrame(patients_data[patients_data['Current Status'] == 'Deceased']['Gender'].values,columns= ['Gender'])
temp = temp.fillna('NA')
temp.loc[(temp['Gender'] == 'Non-Binary'),'Gender'] = np.nan
gender_data_deseased = temp.groupby('Gender').size().reset_index(name = 'size')
fig = go.Figure().add_trace(go.Pie(values=gender_data_deseased['size'],labels= gender_data_deseased['Gender'],
hole = .5,text=gender_data_deseased['Gender']))
fig.update_layout(margin=dict(l=0,r=20,b=20,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'Percentage of cases resulted in death by gender',font_size=20,
xaxis_title="Country",
yaxis_title="Rate")
fig.show()
Testing Analysis¶
state_tests_data.fillna(0,inplace = True)
state_tests_data['Total Tested'] = state_tests_data['Total Tested'].astype('int')
state_tests_data['Positive'] = state_tests_data['Positive'].astype('int')
state_tests_data['Tests per million'] = state_tests_data['Tests per million'].astype('int')
state_tests_data.loc[(state_tests_data['State'] == 'Dadra and Nagar Haveli and Daman and Diu'),'State'] = 'Dadar and Daman'
con = state_data[['State','Confirmed','Deaths']]
tests = state_tests_data.groupby('State')[['Total Tested','Tests per million']].max().reset_index()
# tests = tests.drop('index',axis = 1)
tests = tests.merge(con,how = 'left', on = 'State')
tests.fillna(0,inplace = True)
tests['mortality rate'] = (tests['Deaths']/tests['Confirmed'])*100
tests['spread rate'] = (tests['Confirmed']/tests['Total Tested'])*100
tests['Confirmed'] = tests['Confirmed'].astype('int')
tests['Deaths'] = tests['Deaths'].astype('int')
tests.fillna(0,inplace = True)
tests_named = tests.rename(columns={'Total Tested':'| Total Tested |','Confirmed':'| Confirmed |',
'Tests per million':'| Tests Per Million |',
'Deaths':'| Deaths |','mortality rate':'| Mortality Rate |',
'spread rate':'| Spread Rate |'})
tests_named = tests_named.set_index('State')
tests_named.sort_values('| Total Tested |', ascending= False).style\
.background_gradient(cmap='Blues',subset=["| Total Tested |"])\
.background_gradient(cmap='Greens',subset=["| Tests Per Million |"])\
.background_gradient(cmap='binary',subset=["| Confirmed |"])\
.background_gradient(cmap='Reds',subset=["| Deaths |"])\
.background_gradient(cmap='YlOrBr',subset=["| Mortality Rate |"])\
.background_gradient(cmap='Purples',subset=["| Spread Rate |"])
tests = tests.sort_values('Total Tested')
fig = go.Figure()
fig.add_trace(go.Bar(y=tests['State'],x=tests['Total Tested'],
orientation = 'h',text = tests['Total Tested'],textposition = 'outside',
name='Confirmed',marker= { 'color': 'rgb(47,138,0)'}))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 800,
legend=dict(x=.01,y=.98),
title_text = 'Comparing number of tests done by states',font_size=15,
yaxis_title="State",
xaxis_title="Number of tests (in million)",)
fig.layout.hovermode = 'y'
fig.show()
fig = go.Figure()
fig.add_trace(go.Bar(y=tests['State'],x=tests['Tests per million'],
orientation='h',text = tests['Tests per million'],textposition = 'outside',
name='Confirmed',marker= { 'color': 'rgb(105,67,144)'}))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 800,
legend=dict(x=.01,y=.98),
title_text = 'Number of tests per million of population in states',font_size=15,
yaxis_title="State",
xaxis_title="Number of tests per million",)
fig.layout.hovermode = 'y'
fig.show()
fig = go.Figure()
fig.add_trace(go.Bar(y=tests['State'],x=tests['Tests per million'],orientation = 'h',
name = 'Tests per million',marker= { 'color': 'rgb(105,67,144)'}))
fig.add_trace(go.Bar(y=tests['State'],x=tests['Total Tested'],orientation = 'h',
text = tests['Tests per million'],textposition = 'outside',
name = 'Total Tests',marker= { 'color': 'rgb(47,138,0)'}))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 800,
legend=dict(x=.825,y=.1),
title_text = 'No. of tests and No. of test per million comparison',font_size=15,
yaxis_title="State",
xaxis_title="Number of tests",
barmode='stack')
fig.layout.hovermode = 'y'
fig.show()
Mortality and Spread Rates Comparison¶
national_case_data = case_data[case_data['date'] > '2020-03-25'].groupby('date').sum().reset_index()
national_case_data['mortality Rate'] = (national_case_data['deaths']/national_case_data['confirmed'])*100
fig = go.Figure()
fig.add_trace(go.Scatter(x=national_case_data['date'],y=national_case_data['mortality Rate'],mode='lines+markers',
fill='tozeroy',name = 'Total Tests',marker= { 'color': '#B87625','size' : 10},line=dict( width=4)))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'National Mortality Rate',font_size=15,
xaxis_title="Date",
yaxis_title="Rate")
fig.layout.hovermode = 'x'
fig.show()
tests_mortality = tests[tests['mortality rate']>0].sort_values('mortality rate',ascending = True)
fig = go.Figure()
fig.add_trace(go.Bar(y=tests_mortality['State'],x=tests_mortality['mortality rate'],
orientation = 'h',text = tests['mortality rate'],textposition = 'outside',
name = 'Total Tests',marker= { 'color': '#D38B2C'}))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
title_text = 'Mortality Rate (No. of confirmed / No. of deaths)',font_size=15,
xaxis_title="Country",
yaxis_title="Rate")
fig.layout.hovermode = 'y'
fig.show()
tests_spread = tests.sort_values('spread rate',ascending = True).head(40)
fig = go.Figure()
fig.add_trace(go.Bar(y=tests_spread['State'],x=tests_spread['spread rate'],orientation = 'h',
text = tests['spread rate'],textposition = 'outside',
name = 'Total Tests',marker= { 'color': '#B325B8'}))
fig.update_layout(margin=dict(l=0,r=20,b=0,t=60,pad=0),
paper_bgcolor="white",height= 800,
legend=dict(x=.825,y=.98),
title_text = 'Spreat Rate (No. of Tests / No. of confirmed cases)',font_size=15,
xaxis_title="Country",
yaxis_title="Rate")
fig.layout.hovermode = 'x'
fig.show()
108 countries where India supplied HCQ/Paracetamol¶
HTML(''' <iframe title="[Covid-19: India is supplying HCQ, Paracetamol to 108 countries ]" aria-label="Map" id="datawrapper-chart-kqJFM" src="https://datawrapper.dwcdn.net/kqJFM/1/" scrolling="no" frameborder="0" style="width: 0; min-width: 100% !important; border: none;" height="537"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(a){if(void 0!==a.data["datawrapper-height"])for(var e in a.data["datawrapper-height"]){var t=document.getElementById("datawrapper-chart-"+e)||document.querySelector("iframe[src*='"+e+"']");t&&(t.style.height=a.data["datawrapper-height"][e]+"px")}}))}();
</script> ''')
Containment Zones¶
zones[zones['Zone'] == 'Red Zone'][['District','State','Zone']].reset_index().drop('index',axis=1).style.set_properties(**{'background-color': 'white',
'color': 'black',
'border-color': 'white'})
zones[zones['Zone'] == 'Orange Zone'][['District','State','Zone']].reset_index().drop('index',axis=1).style.set_properties(**{'background-color': 'white',
'color': 'black',
'border-color': 'white'})
zones[zones['Zone'] == 'Green Zone'][['District','State','Zone']].reset_index().drop('index',axis=1).style.set_properties(**{'background-color': 'white',
'color': 'black',
'border-color': 'white'})
zones_data = zones[['State','Zone','District']]
zones_data = zones_data.dropna()
fig = px.treemap(zones_data, path=['State','Zone','District'], color='Zone',
color_discrete_map={'(?)':'#6A6E68','Green Zone':'#00FF00', 'Red Zone':'#E12828', 'Orange Zone':'#C96800'},
hover_data=['Zone'])
fig.update_layout(margin=dict(l=0,r=0,b=0,t=0,pad=0),
paper_bgcolor="white",height= 600,
legend=dict(x=.825,y=.98),
font_size=20)
fig.show()
Prediction¶
Logistic curve is used to predict future behavior of the total number of confirmed cases
- L (the maximum number of confirmed cases) = 250000 taken from the US example (this is from long time obsolete now)
- k (growth rate) = 0.25 approximated value from most of the countries
- x0 (the day of the inflexion) = 80 approximated
The curve being:
y=L/(1+e−k(x−x0))+1¶
import scipy
def logistic(x, L, k, x0):
return L / (1 + np.exp(-k * (x - x0))) + 1
d_df = case_data.groupby('date').max().reset_index()[['date','confirmed']]
def plot_logistic_fit_data(d_df, title, p0):
d_df = d_df.sort_values(by=['date'], ascending=True)
d_df['x'] = np.arange(len(d_df)) + 1
d_df['y'] = d_df['confirmed']
x = d_df['x']
y = d_df['y']
c2 = scipy.optimize.curve_fit(logistic, x, y, p0=p0 )
#y = logistic(x, L, k, x0)
popt, pcov = c2
x = range(1,d_df.shape[0] + int(popt[2]))
y_fit = logistic(x, *popt)
p_df = pd.DataFrame()
p_df['x'] = x
p_df['y'] = y_fit.astype(int)
print("Predicted L (the maximum number of confirmed cases): " + str(int(popt[0])))
print("Predicted k (growth rate): " + str(float(popt[1])))
print("Predicted x0 (the day of the inflexion): " + str(int(popt[2])) + "")
x0 = int(popt[2])
traceC = go.Scatter(x=d_df['x'], y=d_df['y'],name="Confirmed",marker=dict(color="Red"),
mode = "markers+lines",text=d_df['confirmed'],)
traceP = go.Scatter(x=p_df['x'], y=p_df['y'],name="Predicted",marker=dict(color="blue"),
mode = "lines",text=p_df['y'],)
trace_x0 = go.Scatter(x = [x0, x0], y = [0, p_df.loc[p_df['x']==x0,'y'].values[0]],
name = "X0 - Inflexion point",marker=dict(color="black"),
mode = "lines",text = "X0 - Inflexion point")
data = [traceC, traceP, trace_x0]
layout = dict(title = 'Cumulative Conformed cases and logistic curve projection',
xaxis = dict(title = 'Day since first case', showticklabels=True),
yaxis = dict(title = 'Number of cases'),
hovermode = 'x',plot_bgcolor='rgb(275, 270, 273)')
fig = go.Figure(dict(data=data, layout=layout))
fig.show()
L = 250000
k = 0.25
x0 = 100
p0 = (L, k, x0)
plot_logistic_fit_data(d_df, 'India',p0)
import datetime
import scipy
p0 = (0,0)
def plot_exponential_fit_data(d_df, title, delta, p0):
d_df = d_df.sort_values(by=['date'], ascending=True)
d_df['x'] = np.arange(len(d_df)) + 1
d_df['y'] = d_df['confirmed']
x = d_df['x'][:-delta]
y = d_df['y'][:-delta]
c2 = scipy.optimize.curve_fit(lambda t,a,b: a*np.exp(b*t), x, y, p0=p0)
A, B = c2[0]
print(f'(y = Ae^(Bx)) A: {A}, B: {B}')
x = range(1,d_df.shape[0] + 1)
y_fit = A * np.exp(B * x)
traceC = go.Scatter( x=d_df['x'][:-delta], y=d_df['y'][:-delta],name="Confirmed (included for fit)",
marker=dict(color="Red"),mode = "markers+lines",text=d_df['confirmed'])
traceV = go.Scatter(x=d_df['x'][-delta-1:], y=d_df['y'][-delta-1:],name="Confirmed (validation)",
marker=dict(color="blue"),mode = "markers+lines",text=d_df['confirmed'])
traceP = go.Scatter(x=np.array(x), y=y_fit,name="Projected values (fit curve)",marker=dict(color="green"),
mode = "lines",text=y_fit,
)
data = [traceC, traceV, traceP]
layout = dict(title = 'Cumulative Conformed cases and exponential curve projection',
xaxis = dict(title = 'Day since first case', showticklabels=True),
yaxis = dict(title = 'Number of cases'),plot_bgcolor='rgb(275, 270, 273)',
hovermode = 'x')
fig = go.Figure(dict(data=data, layout=layout))
fig.show()
p0 = (40, 0.2)
plot_exponential_fit_data(d_df, 'I', 7, p0)