Covid-19 Cases and Deaths analysis around the world

Shubh Lavti, IIT Gandhinagar, sunil.sl@iitgn.ac.in

Chetan Kishore, IIT Gandhinagar, chetan.k@iitgn.ac.in

Introduction

The 21st century saw its first worldwide pandemic in the form of Covid-19. The pandemic shook the entire world from its onset in December 2019. It caused millions of deaths across the globe. The effects of the pandemic are still visible in various parts of the world. With millions of cases and deaths, it becomes crucial to study the growth, spread, and decline patterns of cases. These studies also help in predicting the future of the covid waves. Hence, WHO has maintained data on Covid-19 spread in different regions of the world. With the help of data from WHO, we can also determine vaccination distribution status, fatality and recovery rates in different areas of the world.

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import plotly.express as px
import plotly
plotly.offline.init_notebook_mode()
import warnings
warnings.filterwarnings('ignore')

df = pd.read_csv("./data/WHO-COVID-19-global-table-data.csv")
df = df.drop(df[df.iloc[:,3]==0].index)
df

	Name	WHO Region	Cases - cumulative total	Cases - cumulative total per 100000 population	Cases - newly reported in last 7 days	Cases - newly reported in last 7 days per 100000 population	Cases - newly reported in last 24 hours	Deaths - cumulative total	Deaths - cumulative total per 100000 population	Deaths - newly reported in last 7 days	Deaths - newly reported in last 7 days per 100000 population	Deaths - newly reported in last 24 hours
0	Global	abs	493392853	6329.980092	7968332	102.229659	1157995	6165833	79.104511	23607	0.302866	3930
1	United States of America	Americas	79501007	24018.239000	207355	62.645000	35049	975540	294.723000	3657	1.105000	610
2	India	South-East Asia	43031958	3118.248000	7518	0.545000	1033	521530	37.792000	401	0.029000	43
3	Brazil	Americas	30040129	14132.580000	157732	74.206000	27331	660528	310.750000	1287	0.605000	216
4	France	Europe	25623682	39397.193000	926709	1424.843000	161950	139706	214.802000	755	1.161000	128
...	...	...	...	...	...	...	...	...	...	...	...	...
226	Falkland Islands (Malvinas)	Americas	125	3588.860000	2	57.422000	0	0	0.000000	0	0.000000	0
227	Holy See	Europe	26	3213.844000	0	0.000000	0	0	0.000000	0	0.000000	0
228	Niue	Western Pacific	7	432.633000	0	0.000000	0	0	0.000000	0	0.000000	0
229	Marshall Islands	Western Pacific	4	6.757000	0	0.000000	0	0	0.000000	0	0.000000	0
230	Nauru	Western Pacific	3	27.691000	3	27.691000	1	0	0.000000	0	0.000000	0

231 rows × 12 columns

From the data provided, we understand that there were 231 countries which were affected by Covid-19. Few data's which we require for the study are of population in different regions, so that we can compare the spread based on populations in different regions. Also for some countries, data is not available as they do not come in the WHO region. The solution to this problem, is that we can assume that there were 0 cases in those countries, as these left out countries are small and do not contribute majorly to the spread of the pandemic.

Col = ["Cases - cumulative total per 100000 population","Cases - newly reported in last 7 days per 100000 population","Deaths - cumulative total per 100000 population","Deaths - newly reported in last 7 days per 100000 population"]
# required_columns = ["Name", "WHO Region","Cases - cumulative total", "Cases - newly reported in last 7 days", "Cases - newly reported in last 24 hours","Deaths - cumulative total","Deaths - newly reported in last 7 days", "Deaths - newly reported in last 24 hours"]
covid_df = df.drop(Col, axis=1)
# covid_df = covid_df.sort_values("Cases - cumulative total")

new_df = covid_df
uniqueValsList = list(set(df["WHO Region"]))

Continents = {}
for i in uniqueValsList:
    Continents[i] = []
for i in range(len(df)):
    Continents[df["WHO Region"][i]].append(df["Name"][i]) 
del Continents['abs']
del Continents['Other']
listconti = list(Continents.keys())
Countries = []

for i in Continents:
    Countries.append(len(Continents[i]))
Countries
fig = plt.figure(figsize = (20,5))
plt.bar(listconti,Countries,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Countries")
plt.title("Number of Countries in each Region")

myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Countries in each Region")
ax=plt.figure()
ax.set_facecolor('blue')
plt.pie(Countries,labels= listconti,startangle=90,explode = myexplode,shadow=True)
plt.show()

It is important to know about the number of countries in a region, so as to understand the reason behind the spread between neighbouring countries.

Cases per region

Total Cumulative cases

Cases_per_region = {}
for i in listconti:
    Cases_per_region[i]=0
for i in range(len(covid_df)):
    for j in range(len(listconti)):
        if df["WHO Region"][i] == listconti[j]:
            Cases_per_region[listconti[j]]+=covid_df["Cases - cumulative total"][i]
            break

x=list(Cases_per_region.keys())
y=list(Cases_per_region.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Cases")
# plt.title("Number of Cases in each Region")
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Cases in each Region")
ax=plt.figure()
ax.set_facecolor('blue')
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)
plt.show()

From the above charts we can see that Covid's affect is completely based on regions and not on number of countries in a region (Africa has a lot of countries, but was least affected by Covid). There is no uniform spread of covid throughout the different regions. The most adverse effect of Covid was seen in the European region and the American region. A determining factor for this type of spread in the European region (even though it has comparatively less population) is because European countries are totally based on tourism and there is a lot of immigration factors. Hence, the spread being the highest. Also, we can say that Covid did not spread uniformly. It was based on environmental factors as well(African region not being affected much)

Cases_per_region_7 = {}
for i in listconti:
    Cases_per_region_7[i]=0
for i in range(len(covid_df)):
    for j in range(len(listconti)):
        if df["WHO Region"][i] == listconti[j]:
            Cases_per_region_7[listconti[j]]+=covid_df["Cases - newly reported in last 7 days"][i]
            break

x=list(Cases_per_region_7.keys())
y=list(Cases_per_region_7.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Cases")
# plt.title("Number of Cases in each Region")
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Cases in each Region in last 7 days")
ax=plt.figure()
ax.set_facecolor('blue')
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)

plt.show()

From the day of release of the data, we observe that Europe is seeing the highest increase in Covid cases in past 7 days. Again, this is credited to a lot of international movement in the European countries. Another factor which may seem to happen is about the knowledge of the cases. In many regions, Covid cases were not registered (in underdeveloped countries) and so in the developed countries, such as the European countries, most of the cases were registered and tended to.

Case_per_region_24 = {}
for i in listconti:
    Case_per_region_24[i]=0
for i in range(len(covid_df)):
    for j in range(len(listconti)):
        if df["WHO Region"][i] == listconti[j]:
            Case_per_region_24[listconti[j]]+=covid_df["Cases - newly reported in last 24 hours"][i]
            break

x=list(Case_per_region_24.keys())
y=list(Case_per_region_24.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Cases")
# plt.title("Number of Cases in each Region")
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Cases in each Region in last 24 hours")
ax=plt.figure()
ax.set_facecolor('blue')
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)
plt.show()

This chart shows that the past 24 hours trend is almost similar to the 7 day trend. And it should be the case, as the trends of the past 24 hours also contribute to the past 7 days trend.

T = {}
for i in Cases_per_region:
    T[i] = [Cases_per_region_7[i],Case_per_region_24[i]]

X = []
y1 = []
y2 = []
cnt = 0
for i in Case_per_region_24:
    
    X.append(cnt+2)
    y1.append(Cases_per_region_7[i])
    y2.append(Case_per_region_24[i])
    cnt+=2
labels = list(Case_per_region_24.keys())

fig = plt.figure(figsize=(20,5))
plt.ylabel("Cases in million")
plt.xlabel("WHO Region")
plt.title("Trend of Covid-19 Cases")
plt.plot(X,y1,label = "Last 7 Days")
plt.plot(X,y2,label = "Last 24 Hours")
plt.xticks(X,labels)
plt.legend()
plt.show()

y3 = []

for i in range(len(y1)):
    y3.append(y2[i]/y1[i])

fig = plt.figure(figsize=(20,5))
plt.plot(X,y3,label = "Ratio of Case in last 24 hours to last 7 days")
plt.xlabel("WHO Region")
plt.ylabel("Ratio")
plt.xticks(X,labels)
plt.legend()
plt.show()
y3

[0.15027388861779462,
 0.1372273576070679,
 0.16017252249332714,
 0.14464966780054914,
 0.09321024419297201,
 0.144270942744439]

From this we can say that the rise in case in last 24 hours is very less in Africa while it has been extremely high for the American and European Countries. So we can predict that the number of cases will spike in the American and European countries for the near future.

Now let us move to the deaths due to Covid-19.

Deaths_perRegion = {}
Deaths_perRegion_7 = {}
Deaths_perRegion_24 = {}
for i in listconti:
    Deaths_perRegion[i]=0
    Deaths_perRegion_7[i]=0
    Deaths_perRegion_24[i]=0
for i in range(len(covid_df)):
    for j in range(len(listconti)):
        if df["WHO Region"][i] == listconti[j]:
            Deaths_perRegion[listconti[j]]+=covid_df["Deaths - cumulative total"][i]
            Deaths_perRegion_7[listconti[j]]+=covid_df["Deaths - newly reported in last 7 days"][i]
            Deaths_perRegion_24[listconti[j]]+=covid_df["Deaths - newly reported in last 24 hours"][i]
            break

x=list(Deaths_perRegion.keys())
y=list(Deaths_perRegion.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Deaths")
plt.title("Number of Deaths in each Region")
plt.show()
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Deaths in each Region")
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)
fig.set_facecolor('blue')
plt.show()


x=list(Deaths_perRegion_7.keys())
y=list(Deaths_perRegion_7.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Deaths")
plt.title("Number of Deaths in each Region in last 7 days")
plt.show()
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Deaths in each Region in last 7 days")
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)
fig.set_facecolor('blue')
plt.show()


x=list(Deaths_perRegion_24.keys())
y=list(Deaths_perRegion_24.values())


fig = plt.figure(figsize = (20,5))
plt.bar(x,y,color='blue',width = 0.2)
plt.xlabel("WHO Region")
plt.ylabel("Deaths")
plt.title("Number of Deaths in each Region in last 24 hours")
plt.show()
myexplode = [0.1,0.1,0.1,0.1,0.1,0.1]
plt.title("Number of Deaths in each Region in last 24 hours")
plt.pie(y,labels= x,startangle=90,explode = myexplode,shadow=True)
fig.set_facecolor('blue')
plt.show()

All over the world there have been millions of deaths, thus understanding the above three charts on death cases is of utmost importance.

The chart on overall death cases directly implies that even if the number of cases were way higher for European region, the number of death cases were comparatively lesser. We can infer from the chart that with the maximum spread of Covid in the European region. They have brought down the death tolls significantly with help of improved medical services. On similar grounds we can say that the other regions such as Africa and South East Asia, these regions are lacking in medical services.

The last two charts, refering to the death tolls in past 7 days and past 24 hours implies that all over the world, countries were trying to control the number of death tolls. The Western Pacific region could not control the death tolls, as we can infer from the charts.

x = []
y1 = []
y2 = []
y3 = []
cnt = 0
for i in listconti:
    x.append(cnt+2)
    y1.append(Case_per_region_24[i])
    y2.append(Deaths_perRegion_24[i])
    a = 100*Deaths_perRegion_24[i]/Case_per_region_24[i]
    y3.append(a)
    cnt+=2

fig = plt.figure(figsize = (20,5))
plt.plot(x,y3)
plt.xlabel("WHO Region")
plt.ylabel("Fatality Rate in percent")
plt.title("Deaths and Cases per Region")
plt.xticks(x,labels)
plt.show()

The above graph directly relates to the medical services and vaccination status of diffierent regions across the globe. Countries in American region as well as in Eastern Mediterranean region have suffered a lot. It may be that the disease's effect may be critical in these regions, whereas it had very less impact in Africa.

covid_df = covid_df.sort_values("Cases - cumulative total")
top_five = covid_df[len(covid_df)-6:len(covid_df)-1]
top_five = top_five.sort_values("Cases - cumulative total")
top_five

	Name	WHO Region	Cases - cumulative total	Cases - newly reported in last 7 days	Cases - newly reported in last 24 hours	Deaths - cumulative total	Deaths - newly reported in last 7 days	Deaths - newly reported in last 24 hours
5	Germany	Europe	22265788	1161279	201729	131036	1645	328
4	France	Europe	25623682	926709	161950	139706	755	128
3	Brazil	Americas	30040129	157732	27331	660528	1287	216
2	India	South-East Asia	43031958	7518	1033	521530	401	43
1	United States of America	Americas	79501007	207355	35049	975540	3657	610

top_five_country = list(top_five["Name"])
top_five_country

['Germany', 'France', 'Brazil', 'India', 'United States of America']

Although Europe has very high number of cases in Recent times but the fatality rate is still pretty less as compared to other WHO Regions. The Fatality rate is highest in American Continent. Thus the distribution of vaccine can be proposed on the basis of this rates.

date_df = pd.read_csv("./data/WHO-COVID-19-global-data.csv")
date_df

	Date_reported	Country_code	Country	WHO_region	New_cases	Cumulative_cases	New_deaths	Cumulative_deaths
0	2020-01-03	AF	Afghanistan	EMRO	0	0	0	0
1	2020-01-04	AF	Afghanistan	EMRO	0	0	0	0
2	2020-01-05	AF	Afghanistan	EMRO	0	0	0	0
3	2020-01-06	AF	Afghanistan	EMRO	0	0	0	0
4	2020-01-07	AF	Afghanistan	EMRO	0	0	0	0
...	...	...	...	...	...	...	...	...
195757	2022-04-03	ZW	Zimbabwe	AFRO	195	246481	2	5446
195758	2022-04-04	ZW	Zimbabwe	AFRO	44	246525	0	5446
195759	2022-04-05	ZW	Zimbabwe	AFRO	87	246612	5	5451
195760	2022-04-06	ZW	Zimbabwe	AFRO	132	246744	0	5451
195761	2022-04-07	ZW	Zimbabwe	AFRO	59	246803	2	5453

195762 rows × 8 columns

Country = {}
dates = []
for i in range(len(date_df)):
    if date_df["Country"][i] in Country:
        Country[date_df["Country"][i]].append(date_df["New_cases"][i])
    else:
        Country[date_df["Country"][i]] = [date_df["New_cases"][i]]

    if date_df["Date_reported"][i] not in dates:
        dates.append(date_df["Date_reported"][i])
# len(dates)
# Country

Months = ["January","February","March","April","May","June","July","August","September","October","November","December"]

for l in top_five_country:
    DATE = []
    x=[]
    cnt = 0
    y=[]
    for i in range(len(dates)//30):
        DATE.append(Months[int(i%12)])
        y.append(0)
    for i in range(len(list(Country[l]))):
        x.append(cnt+1)
        cnt+=1


    fig = plt.figure(figsize = (20,5))
    plt.plot(x,list(Country[l]),color="black")
    xd = []
    flag = 20
    for i in range(7):
        xd.append(flag)
        flag+=120

    DATE = ["January 23, 2020","May 23,2020","September 20,2020","January 18,2020","May 20,2021","September 17,2021","January 15,2022"]
    plt.xticks(xd,DATE)
    plt.title("Covid Cases in "+str(l))
    plt.xlabel("Timeline")
    plt.ylabel("New Cases per Day")
    plt.show()

The graphs of New cases per day shows very rough trend because there is no uniformity in number of cases per day. Everyday, the cases are different and even though it depends on previous days' trends, the data collected is when people go out for their checkups and find out the results. Hence we can expect a rough graph. (Mostly for countries with less population)

In Germany, there was no peak for a long time, but we observe that by the start of 2022, the cases started to rise quickly.

Similar is the case for France.

Brazil has seen a very rough trend throughout.

In India, we can clearly see that there have been three covid spikes, where the 2nd one was the longest one and with maximum cases. Also, here the daily trends are not jacked, because of the large population.

In USA also, there are three clear spikes and that too on a similar timeline as of India. Here, the last spike is the longest.

Country_d = {}
dates = []
for i in range(len(date_df)):
    if date_df["Country"][i] in Country_d:
        Country_d[date_df["Country"][i]].append(date_df["New_deaths"][i])
    else:
        Country_d[date_df["Country"][i]] = [date_df["New_deaths"][i]]

    if date_df["Date_reported"][i] not in dates:
        dates.append(date_df["Date_reported"][i])
len(dates)
Months = ["January","February","March","April","May","June","July","August","September","October","November","December"]

for l in top_five_country:
    DATE = []
    x=[]
    cnt = 0
    y=[]
    for i in range(len(dates)//30):
        DATE.append(Months[int(i%12)])
        y.append(0)
    for i in range(len(list(Country_d[l]))):
        x.append(cnt+1)
        cnt+=1


    fig = plt.figure(figsize = (20,5))
    plt.plot(x,list(Country_d[l]),color="black")
    xd = []
    flag = 20
    for i in range(7):
        xd.append(flag)
        flag+=120

    DATE = ["January 23, 2020","May 23,2020","September 20,2020","January 18,2020","May 20,2021","September 17,2021","January 15,2022"]
    plt.xticks(xd,DATE)
    plt.title("New Deaths in "+str(l))
    plt.xlabel("Timeline")
    plt.ylabel("New Deaths")
    plt.show()

As we can see from the above graphs that the peak in covid-19 cases and deaths are distinct for all the five countries. In India we can observe three waves (September 2020, May 2021, January 2022) in which the most devastating wave was the second wave.

We also observe that between September 2020 to May 2021, there have been maximum deaths throughout the world.

Try out

covid_df = pd.read_csv("./data/owid-covid-data.csv")
covid_df.sample(5)

	iso_code	continent	location	date	total_cases	new_cases	new_cases_smoothed	total_deaths	new_deaths	new_deaths_smoothed	...	female_smokers	male_smokers	handwashing_facilities	hospital_beds_per_thousand	life_expectancy	human_development_index	excess_mortality_cumulative_absolute	excess_mortality_cumulative	excess_mortality	excess_mortality_cumulative_per_million
38133	HRV	Europe	Croatia	2021-02-20	239685.0	431.0	318.000	5420.0	13.0	17.286	...	34.3	39.9	NaN	5.54	78.49	0.851	NaN	NaN	NaN	NaN
148388	KOR	Asia	South Korea	2020-09-29	23812.0	113.0	85.143	413.0	6.0	3.571	...	6.2	40.9	NaN	12.27	83.03	0.916	NaN	NaN	NaN	NaN
20574	BES	North America	Bonaire Sint Eustatius and Saba	2021-11-01	2444.0	9.0	7.714	19.0	0.0	0.000	...	NaN	NaN	NaN	NaN	77.79	NaN	NaN	NaN	NaN	NaN
34079	COL	South America	Colombia	2021-03-10	2285960.0	3588.0	3330.571	60773.0	97.0	98.714	...	4.7	13.5	65.386	1.71	77.29	0.767	NaN	NaN	NaN	NaN
37273	CIV	Africa	Cote d'Ivoire	2020-11-24	21156.0	8.0	21.714	131.0	0.0	0.286	...	NaN	NaN	19.351	NaN	57.78	0.538	NaN	NaN	NaN	NaN

5 rows × 67 columns

required_columns = ["iso_code", "location", "continent", "date", "new_cases_smoothed", "total_cases"]
covid_df = covid_df.dropna(subset = required_columns)

covid_df = covid_df.sort_values("date")

covid_day_df = covid_df[covid_df.date == "2021-01-31"]

fig = px.scatter_geo(covid_day_df, locations="iso_code", color="continent",
                     hover_name="location", size="total_cases",
                     projection="natural earth",title="Total Number of Cases spread across all Continents")

fig.show()

covid_day_df["total_deaths"]=covid_day_df["total_deaths"].fillna(0)

fig = px.scatter_geo(covid_day_df, locations="iso_code", color="continent",
                     hover_name="location", size="total_deaths",
                     projection="natural earth",title="total deaths throughout all continents")

fig.show()

It is an obvious observation that wherever there were huge cases, the deaths have been maximum in those regions. India, USA, France and Brazil are among the most affected countries. The European region has been most affected by covid.