Rogress of Covid 19 Vaccination 1

rogress-of-covid-19-vaccination-1
April 20, 2024
#Analyzing and Predicting Progress of COVID-19 Vaccination
IS833 Business Analytics Project - Team 5

Tuan Cao, Firdaus Faiz, Ansh Patil, Max Renaerts, Amogh Vadjikar, Nancy Vega
0.1 Dataset link:

Source: COVID-19 World Vaccination Progress (Daily and Total Vaccination for COVID-19 in the
World from Our World in Data)
Kaggle link: https://www.kaggle.com/datasets/gpreda/covid-world-vaccination-
progress?select=country_vaccinations.csv
Original Source: COVID-19 Dataset by Our World in Data (link)

• Confirmed cases and deaths: our data comes from the COVID-19 Data Repository by the
Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (JHU). We
discuss how and when JHU collects and publishes this data here. The cases & deaths dataset
is updated daily.
• Hospitalizations and intensive care unit (ICU) admissions: our data is collected from official
sources and collated by Our World in Data. The complete list of country-by-country sources
is available here. Testing for COVID-19: this data is collected by the Our World in Data team
from official reports; you can find further details in our post on COVID-19 testing, including
our checklist of questions to understand testing data, information on geographical and tem-
poral coverage, and detailed country-by-country source information. The testing dataset is
updated around twice a week. Vaccinations against COVID-19: this data is collected by the
Our World in Data team from official reports. Other variables: this data is collected from a
variety of sources (United Nations, World Bank, Global Burden of Disease, Blavatnik School
of Government, etc.). More information is available in our codebook.
0.2 Introduction & Project Objective

Since late 2019, we have been living in a global pandemic known as COVID-19. Thanks to the
vaccine, we were able to survive the pandemic. Our notebook will guide you through the distri-
bution of the different vaccine brands over time, provide a geographical overview and predict the
vaccination rate.
1
0.3 Problem:
Our goal is to track COVID-19 vaccine adoption in the World. We will analyze the rate of people
getting vaccinated per day in each country and the types of vaccine used. Additionally using the
historical data, we hope to predict the progress of vaccination and immunization status in these
countries.
0.4 Data importation

[ ]: import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
# Enlarge charts produced

plt.rcParams["figure.figsize"] = (20,10)
[ ]: from google.colab import drive

drive.mount('/content/drive')
Mounted at /content/drive
[ ]: # Upload file from Kaggle to be read here

# pd.read_csv("drive/MyDrive/Business Analytics_IS883_Team Project/
↪country_vaccinations.csv")
country_vaccine = pd.read_csv("drive/MyDrive/Business Analytics_IS833_Team␣

↪Project/country_vaccinations.csv")
company_vaccine = pd.read_csv("drive/MyDrive/Business Analytics_IS833_Team␣

↪Project/country_vaccinations_by_manufacturer.csv")
country_vaccine.head()
[ ]: country iso_code date total_vaccinations people_vaccinated \

0 Afghanistan AFG 22-02-21 0.0 0.0
1 Afghanistan AFG 23-02-21 NaN NaN
people_fully_vaccinated daily_vaccinations_raw daily_vaccinations \

0 NaN NaN NaN
1 NaN NaN 1367.0
2 NaN NaN 1367.0
3 NaN NaN 1367.0
4 NaN NaN 1367.0
total_vaccinations_per_hundred people_vaccinated_per_hundred \
0 0.0 0.0
2
1 NaN NaN
2 NaN NaN
3 NaN NaN
4 NaN NaN
people_fully_vaccinated_per_hundred daily_vaccinations_per_million \
0 NaN NaN
1 NaN 34.0
2 NaN 34.0
3 NaN 34.0
4 NaN 34.0
vaccines \
0 Johnson&Johnson, Oxford/AstraZeneca, Pfizer/Bi…
source_name source_website
0 World Health Organization https://covid19.who.int/
[ ]: # Table 2
company_vaccine.head()
[ ]: location date vaccine total_vaccinations

0 Argentina 2020-12-29 Moderna 2
1 Argentina 2020-12-29 Oxford/AstraZeneca 3
2 Argentina 2020-12-29 Sinopharm/Beijing 1
3 Argentina 2020-12-29 Sputnik V 20481
4 Argentina 2020-12-30 Moderna 2
0.5 Data Categories

The data (country vaccinations) contains the following information:
• Country - this is the country for which the vaccination information is provided;
• Country ISO Code - ISO code for the country;
• Date - date for the data entry; for some of the dates we have only the daily vaccinations, for
others, only the (cumulative) total;
• Total number of vaccinations - this is the absolute number of total immunizations in the
country;
• Total number of people vaccinated - a person, depending on the immunization scheme,
will receive one or more (typically 2) vaccines; at a certain moment, the number of vaccination
3
might be larger than the number of people;
• Total number of people fully vaccinated - this is the number of people that received the
entire set of immunization according to the immunization scheme (typically 2); at a certain
moment in time, there might be a certain number of people that received one vaccine and
another number (smaller) of people that received all vaccines in the scheme;
• Daily vaccinations (raw) - for a certain data entry, the number of vaccination for that
date/country;
• Daily vaccinations - for a certain data entry, the number of vaccination for that
date/country;
• Total vaccinations per hundred - ratio (in percent) between vaccination number and total
population up to the date in the country;
• Total number of people vaccinated per hundred - ratio (in percent) between population
immunized and total population up to the date in the country;
• Total number of people fully vaccinated per hundred - ratio (in percent) between
population fully immunized and total population up to the date in the country;
• Number of vaccinations per day - number of daily vaccination for that day and country;
• Daily vaccinations per million - ratio (in ppm) between vaccination number and total
population for the current date in the country;
• Vaccines used in the country - total number of vaccines used in the country (up to date);
• Source name - source of the information (national authority, international organization,
local organization etc.);
• Source website - website of the source of information;
0.6 Data Cleaning and Validation Checks

[ ]: # Sample code to reformat DMY to YMD
country_vaccine['date'] = pd.to_datetime(country_vaccine['date'],␣
↪format='%d-%m-%y').dt.strftime("%Y-%m-%d")
# Display country_vaccine with new date format

country_vaccine.head()
[ ]: country iso_code date total_vaccinations people_vaccinated \

0 Afghanistan AFG 2021-02-22 0.0 0.0
people_fully_vaccinated daily_vaccinations_raw daily_vaccinations \

0 NaN NaN NaN
1 NaN NaN 1367.0
2 NaN NaN 1367.0
3 NaN NaN 1367.0
4 NaN NaN 1367.0
total_vaccinations_per_hundred people_vaccinated_per_hundred \
0 0.0 0.0
4
1 NaN NaN
2 NaN NaN
3 NaN NaN
4 NaN NaN
0 NaN NaN
1 NaN 34.0
2 NaN 34.0
3 NaN 34.0
4 NaN 34.0
vaccines \
source_name source_website
[ ]: ## Sample code to reformat DMY to YMD (^Now changed to pandas function)
# import datetime
# datetime.datetime.strptime("21-12-2008", "%d-%m-%Y").strftime("%Y-%m-%d")
## Reformat date from DMY to YMD

# for i in range(len(country_vaccine)):
# country_vaccine['date'].iloc[i] = country_vaccine['date'].iloc[i][0:6] +␣
↪'20' + country_vaccine['date'].iloc[i][6:]
# country_vaccine['date'].iloc[i] = datetime.datetime.
↪strptime(country_vaccine['date'].iloc[i], "%d-%m-%Y").strftime("%Y-%m-%d")
[ ]: # Display maximum date in timeline

country_vaccine["date"].max()
[ ]: '2022-03-29'
[ ]: # Display minimum date in timeline

country_vaccine["date"].min()
[ ]: '2020-12-02'
5
[ ]: country_vaccine.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 86512 entries, 0 to 86511
Data columns (total 15 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 country 86512 non-null object
1 iso_code 86512 non-null object
2 date 86512 non-null object
3 total_vaccinations 43607 non-null float64
4 people_vaccinated 41294 non-null float64
5 people_fully_vaccinated 38802 non-null float64
6 daily_vaccinations_raw 35362 non-null float64
7 daily_vaccinations 86213 non-null float64
8 total_vaccinations_per_hundred 43607 non-null float64
9 people_vaccinated_per_hundred 41294 non-null float64
10 people_fully_vaccinated_per_hundred 38802 non-null float64
11 daily_vaccinations_per_million 86213 non-null float64
12 vaccines 86512 non-null object
13 source_name 86512 non-null object
14 source_website 86512 non-null object
dtypes: float64(9), object(6)
memory usage: 9.9+ MB
[ ]: company_vaccine.info()
--- ------ -------------- -----
0 location 35623 non-null object
2 vaccine 35623 non-null object
3 total_vaccinations 35623 non-null int64
dtypes: int64(1), object(3)
[ ]: company_vaccine['date'].max()
[ ]: '2022-03-30'
[ ]: company_vaccine['date'].min()
[ ]: '2020-12-04'
6
[ ]: #Pivot Table to have each Vaccine's trend in columns
new_company_vaccine = company_vaccine.pivot(index=['location', 'date'],␣
↪columns='vaccine',values='total_vaccinations')
new_company_vaccine = new_company_vaccine.fillna(0)
new_company_vaccine
[ ]: vaccine CanSino Covaxin Johnson&Johnson Moderna Novavax \

location date
Argentina 2020-12-29 0.0 0.0 0.0 2.0 0.0
2020-12-30 0.0 0.0 0.0 2.0 0.0
2020-12-31 0.0 0.0 0.0 2.0 0.0
2021-01-01 0.0 0.0 0.0 2.0 0.0
2021-01-02 0.0 0.0 0.0 2.0 0.0
… … … … … …
Uruguay 2022-03-25 0.0 0.0 0.0 0.0 0.0
2022-03-26 0.0 0.0 0.0 0.0 0.0
2022-03-27 0.0 0.0 0.0 0.0 0.0
2022-03-28 0.0 0.0 0.0 0.0 0.0
2022-03-29 0.0 0.0 0.0 0.0 0.0
vaccine Oxford/AstraZeneca Pfizer/BioNTech Sinopharm/Beijing \

location date
Argentina 2020-12-29 3.0 0.0 1.0
2020-12-30 3.0 0.0 1.0
2020-12-31 3.0 0.0 1.0
2021-01-01 5.0 0.0 1.0
2021-01-02 6.0 0.0 1.0
… … … …
Uruguay 2022-03-25 90732.0 2473572.0 0.0
2022-03-26 90732.0 2473582.0 0.0
2022-03-27 90732.0 2473582.0 0.0
2022-03-28 90732.0 2475706.0 0.0
2022-03-29 90732.0 2476595.0 0.0
vaccine Sinovac Sputnik V

location date
Argentina 2020-12-29 0.0 20481.0
2020-12-30 0.0 40583.0
2020-12-31 0.0 43388.0
2021-01-01 0.0 43513.0
2021-01-02 0.0 46824.0
… … …
Uruguay 2022-03-25 3247932.0 0.0
2022-03-26 3247932.0 0.0
2022-03-27 3247932.0 0.0
2022-03-28 3247932.0 0.0
2022-03-29 3247932.0 0.0
7
[9046 rows x 10 columns]
[ ]: # India data is not available in the 2nd table␣

↪<country_vaccinations_by_manufacturer.csv> from source
company_test = company_vaccine.
↪groupby(['location','vaccine'])[['total_vaccinations']].max()
company_test.reset_index(inplace=True)
company_test=company_test[company_test['location']=='India']
company_test
[ ]: Empty DataFrame
Columns: [location, vaccine, total_vaccinations]
Index: []
[ ]: #Display the latest values for each Country-Vaccine combination

## Since, the total_vaccinations column is cumulative, max(total_vaccinations)␣
↪will generate the latest value
company_vaccine_aggr = company_vaccine.
↪groupby(['location','vaccine'])[['total_vaccinations']].max()
company_vaccine_aggr.reset_index(inplace=True)
company_vaccine_aggr.sort_values(['total_vaccinations'], ascending=[False],␣
↪inplace=True)
company_aggr = company_vaccine.groupby(['location'])[['total_vaccinations']].
↪max()
company_aggr.reset_index(inplace=True)
company_aggr=company_aggr[company_aggr['location']!='European Union']
#company_vaccine_aggr=company_vaccine_aggr[company_vaccine_aggr['location']!
↪='European Union']
#company_vaccine_aggr
company_aggr.sort_values(['total_vaccinations'], ascending=[False],␣
↪inplace=True)
company_vaccine_aggr=company_aggr.
↪merge(company_vaccine_aggr,on='location',how='inner')
company_vaccine_aggr.
↪columns=['location','country_max','vaccine','total_vaccinations']
company_vaccine_aggr.head()
[ ]: location country_max vaccine total_vaccinations

0 United States 330748673 Pfizer/BioNTech 330748673
1 United States 330748673 Moderna 210341785
2 United States 330748673 Johnson&Johnson 18563903
3 Germany 126041243 Pfizer/BioNTech 126041243
4 Germany 126041243 Moderna 29435818
8
[ ]: #Comment: Min sorting and max sorting
[ ]: df = company_vaccine_aggr
df['Rank']=df['country_max'].rank(ascending=False,method='dense')
df=df[df['Rank']<=10]
df.head()
[ ]: location country_max vaccine total_vaccinations Rank

0 United States 330748673 Pfizer/BioNTech 330748673 1.0
1 United States 330748673 Moderna 210341785 1.0
2 United States 330748673 Johnson&Johnson 18563903 1.0
3 Germany 126041243 Pfizer/BioNTech 126041243 2.0
4 Germany 126041243 Moderna 29435818 2.0
0.7 Exploratory Data Analysis

Two questions to be answered
• Is there a link between analyzed countries and vaccine brands (including availability of each
vaccine)?
• Most popular vaccine used in each country & comparison amongst countries?
The below bar chart shows the top 10 countries with highest total number of Covid-19
vaccinations across the globe. Apparently, Pfizer is the most commonly-used vaccine
(except in Argentina) in these countries.
[ ]: # Plot total vaccinations per country using specific vaccine
sns.set(rc={"figure.figsize":(15, 5)})
barplot = sns.barplot(x='location', y='total_vaccinations',␣

↪data=df,hue="vaccine")
barplot.ticklabel_format(style='plain', axis='y') #change scientific notation␣

↪on x-axis to integer
barplot.tick_params(axis='x', rotation=90) #NV changed to not have overlapping␣

↪country names
barplot.set_ylabel("Total Vaccinations", fontsize = 12) #label x-axis

barplot.set_xlabel("Country", fontsize = 12) #label y-axis
barplot.set_title("Top 10 Countries and Vaccines", fontsize = 15); #set chart␣
↪title
#plt.xlim(0, 650000000) #set same x-axis across other vaccines
9
The below bar chart shows the bottom 10 countries with lowest total number of Covid-19 vacci-
nations across the globe. It can clearly be seen that Pfizer is still the most widely used vaccine in
these countries against Covid except Uruguay.
[ ]: df = company_vaccine_aggr
df['Rank']=df['country_max'].rank(ascending=True,method='dense')
df=df[df['Rank']<=10]
# Plot total vaccinations per country using specific vaccine

barplot = sns.barplot(x='location', y='total_vaccinations',␣

↪data=df,hue="vaccine")
barplot.ticklabel_format(style='plain', axis='y') #change scientific notation␣

barplot.set_ylabel("Total Vaccinations", fontsize = 12) #label x-axis

barplot.set_xlabel("Country", fontsize = 12) #label y-axis
barplot.set_title("Bottom 10 Countries and Vaccines", fontsize = 15); #set␣
↪chart title
10
The following graph maps the top 10 countries using highest number of
Pfizer/BioNTech vaccine. The x-axis shows the total Pfizer/BioNTech shots and y-
axis shows the matching country. We can see that the United States is the biggest
distributor of this brand.
df = company_vaccine_aggr.set_index(['location','vaccine'])
df=df[df.index.get_level_values('vaccine').isin(['Pfizer/BioNTech'])].
↪droplevel(1)
df.sort_values(by='total_vaccinations', ascending=False, inplace=True)

df_top10= df.head(10)
barplot = sns.barplot(y=df_top10.index, x='total_vaccinations', data=df_top10)
barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

barplot.set_ylabel("Country", fontsize = 15) #label y-axis
barplot.set_title("Countries with highest usage of Pfizer/BioNTech Vaccine",␣
↪fontsize = 22); #set chart title
11
The following graph maps the bottom 10 countries using lowest number of Pfizer/BioNTech vaccine.
The x-axis shows the total Pfizer/BioNTech shots and y-axis shows the matching country. We can
see that Liechtenstein is smallest distributor of this brand. It is worth to note that Liechtenstein
is a small nation and there are 332 million more people living in the US than Liechtenstein.
df_tail10= df.tail(10)
barplot = sns.barplot(y=df_tail10.index, x='total_vaccinations', data=df_tail10)


barplot.set_title("Countries with lowest usage of Pfizer/BioNTech Vaccine",␣
12
The following graph maps the top 10 countries using highest number of Moderna
vaccine. The x-axis shows the total Moderna shots and y-axis shows the matching
country. We can see that the United States is still the biggest distributor of this
brand. Comparing the use of Pfizer/BioNTech with Moderna in the US specifically,
Moderna is used 30% less than Pfizer/BioNTech.
df=df[df.index.get_level_values('vaccine').isin(['Moderna'])].droplevel(1)



barplot.set_title("Countries with highest usage of Moderna Vaccine", fontsize =␣
↪22); #set chart title
The following graph maps the bottom 10 countries using lowest number of Moderna vaccine. The
x-axis shows the total Moderna shots and y-axis shows the matching country. We can see that
Iceland is smallest distributor of this brand. It is worth to note that Iceland also has a significantly
smaller population in comparison to the US.
[ ]: df_tail10= df.tail(10)

13
barplot.set_title("Countries with lowest usage of Moderna Vaccine", fontsize =␣
↪22);
The following graph maps the top 10 countries using highest number of Ox-
ford/AstraZeneca vaccine. The x-axis shows the total Oxford/AstraZeneca shots and
y-axis shows the matching country. We can see that the Argentina is the biggest
distributor of this brand, followed by South Korea. The US did not opt for Ox-
ford/AstraZeneca vaccine and hence it is not visible in this list.
df=df[df.index.get_level_values('vaccine').isin(['Oxford/AstraZeneca'])].
↪droplevel(1)



barplot.set_title("Countries with highest usage of Oxford/AstraZeneca Vaccine",␣

14
The following graph maps the bottom 10 countries using lowest number of Oxford/AstraZeneca
vaccine. The x-axis shows the total Oxford/AstraZeneca shots and y-axis shows the matching
country. We can see that Iceland is smallest distributor of this brand - almost negligible.


barplot.set_title("Countries with lowest usage of Oxford/AstraZeneca Vaccine",␣
↪fontsize = 22);
The following graph maps the top 10 countries using highest number of John-
son&Johnson (J&J) vaccine. The x-axis shows the total Johnson&Johnson (J&J)
shots and y-axis shows the matching country. We can see that the United States
is the biggest distributor of this brand, followed by South Africa. However, com-
15
paring the other vaccines used in the US, Johnson&Johnson (J&J) is used less than
Pfizer/BioNTech and Moderna.
df=df[df.index.get_level_values('vaccine').isin(['Johnson&Johnson'])].
↪droplevel(1)



barplot.set_title("Countries with highest usage of Johnson&Johnson Vaccine",␣

The following graph maps the bottom 10 countries using lowest number of Johnson&Johnson (J&J)
vaccine. The x-axis shows the total Johnson&Johnson (J&J) shots and y-axis shows the matching
country. We can see that Leichtenstein, Iceland and Finland are the smallest distributor of this
brand - almost negligible.


16
barplot.set_title("Countries with lowest usage of Johnson&Johnson Vaccine",␣
↪fontsize = 22);
The following graph plots countries using the Sinovac vaccine. The x-axis shows the
total vaccinations and y-axis the matching country. This vaccine is only used relatively
significant in 5 countries with Chile as the biggest one.
df=df[df.index.get_level_values('vaccine').isin(['Sinovac'])].droplevel(1)
barplot = sns.barplot(y=df.index, x='total_vaccinations', data=df)


barplot.set_title("Countries Using Sinovac Vaccine", fontsize = 22); #set␣
↪chart title
17
The following two graphs below give an overview of countries using the Novavax vac-
cine. The x-axis shows the total vaccinations and y-axis the matching country. The
first graph shows the top 10 countries using highest number of Johnson&Johnson
(J&J) vaccine. The second graph is a continuation of the list as there are not many
countries using Novavx in general. Hence, the second graph has some overlaps and
zooms into the countries on the bottom of the list that uses Novavax vaccine. It
can concluded that this vaccine is only used most significantly in 3 countries like South Korea,
Germany, and Italy, with South Korea being the largest distributor.
df=df[df.index.get_level_values('vaccine').isin(['Novavax'])].droplevel(1)

barplot.set_title("Countries with highest usage of Novavax Vaccine", fontsize =␣
↪22); #set chart title
18

barplot.set_title("Countries with lowest usage of Novavax Vaccine", fontsize =␣
↪22);
The following graph plots countries using the Sinopharm vaccine. The x-axis shows
the total vaccinations and y-axis the matching country. This vaccine isn’t largely
distributed over the world. We can see the Argetina, Peru and Nepal are the main
countries using this brand.
19
df=df[df.index.get_level_values('vaccine').isin(['Sinopharm/Beijing'])].
↪droplevel(1)


barplot.set_title("Countries Using Sinopharm/Beijing Vaccine", fontsize = 22); ␣
↪#set chart title
This graph maps out the countries distributing Sputnik, which is mostly used in Ar-
gentina.
df=df[df.index.get_level_values('vaccine').isin(['Sputnik V'])].droplevel(1)


20
barplot.set_title("Countries Using Sputnik V Vaccine", fontsize = 22); #set␣
↪chart title
We can see in this graph that CanSino is only used in 3 countries: Argentina, Chile
and Ecuador
df=df[df.index.get_level_values('vaccine').isin(['CanSino'])].droplevel(1)


barplot.set_title("Countries Using CanSino Vaccine", fontsize = 22); #set␣
↪chart title
21
This graphs shows us that Covaxin is only used in Portugal.
df=df[df.index.get_level_values('vaccine').isin(['Covaxin'])].droplevel(1)


barplot.set_title("Countries Using Covaxin Vaccine", fontsize = 22); #set␣
↪chart title
22
[ ]: company_vaccine.info()
--- ------ -------------- -----
0 location 35623 non-null object
2 vaccine 35623 non-null object
3 total_vaccinations 35623 non-null int64
dtypes: int64(1), object(3)
[ ]: new_company_vaccine.info()
MultiIndex: 9046 entries, ('Argentina', '2020-12-29') to ('Uruguay',
'2022-03-29')
--- ------ -------------- -----
0 CanSino 9046 non-null float64
1 Covaxin 9046 non-null float64
2 Johnson&Johnson 9046 non-null float64
3 Moderna 9046 non-null float64
4 Novavax 9046 non-null float64
5 Oxford/AstraZeneca 9046 non-null float64
6 Pfizer/BioNTech 9046 non-null float64
7 Sinopharm/Beijing 9046 non-null float64
8 Sinovac 9046 non-null float64
9 Sputnik V 9046 non-null float64
dtypes: float64(10)
memory usage: 754.6+ KB
[ ]: # Plot graph of daily vaccinations vs time for a specific country
import matplotlib.ticker as ticker
graph = country_vaccine[country_vaccine.country == 'United States'].plot(x =␣

↪"date", y = "daily_vaccinations_per_million")
graph.xaxis.set_major_locator(ticker.LinearLocator(10)) #adjust x-axis label to␣

↪make it readable
graph.ticklabel_format(style='plain', axis='y') #change scientific notation on␣

↪y-axis to integer
graph.set_xlabel("Date", fontsize = 15) #label x-axis
23
graph.set_ylabel("Daily Vaccination", fontsize = 15) #label y-axis
graph.set_title("Daily Vaccination Trends over Time", fontsize = 22) #set␣
↪chart title
[ ]: Text(0.5, 1.0, 'Daily Vaccination Trends over Time')
Here above is a line graph depicting the Daily Vaccination Trends over time, starting December
2021 through January 2022.
After cleaning up the data, we wanted to understand the trends of daily vaccine rates overtime as
we noticed a quick decline among populations. It was an obvious prediction of decrease in speed
given the number of people whom already received their initial shots, but also political climate
effects. There was a continous global vaccine hesitancy that slowed down most countries’ progress
in vaccination rates. Thus causing more federal government rules and regulations regarding vaccine
requirements, which can explain the downward trend you can see within this graph.
Eventually, we can expect that daily vaccinations to hit 0, once everyone becomes fully vaccinated.
Once we have a general understanding of the daily vaccinations trend over time, we wanted to
zoom into specific countries in the top 10 and bottom 100 with regards to GDP ranking to identify
similarities and differences.
As expected, countries with high GDP like China, India, United States, and Japan dominates other
countries considered in this list when it comes to daily vaccination.
[ ]: # Plot graph of daily vaccinations vs time for multiple countries for comparison
# Chosen top 5 countries and bottom 100 countries by GDP rank

chart = sns.lineplot(data=country_vaccine[country_vaccine.country.isin(['United␣
↪States', 'Japan', 'Germany', 'United␣
↪Kingdom','Venezuela','Rwanda','Paraguay','Cambodia','India','China'])],
x='date', y='daily_vaccinations', hue='country')
24
chart.xaxis.set_major_locator(ticker.LinearLocator(10)) #adjust x-axis label to␣
↪make it readable
chart.ticklabel_format(style='plain', axis='y') #change scientific notation on␣

chart.set_xlabel("Date", fontsize = 15) #label x-axis

chart.set_ylabel("Daily Vaccination", fontsize = 15) #label y-axis
chart.set_title("Daily Vaccination Trends over Time", fontsize = 22) #set␣
↪chart title
[ ]: Text(0.5, 1.0, 'Daily Vaccination Trends over Time')
The line chart above on “daily vaccinations” gave accurate results based on the volume. However,
we chose to modify the visualization to “daily vaccination per million” to better understand the
coverage of these vaccines within specific countries as can be seen in the chart below.
[ ]: # Plot graph of daily vaccinations vs time for multiple countries for comparison
# Chosen top 5 countries and bottom 100 countries by GDP rank

chart = sns.lineplot(data=country_vaccine[country_vaccine.country.isin(['United␣
↪States','China', 'Japan', 'Germany', 'United␣
↪Kingdom','Venezuela','Rwanda','Paraguay','Cambodia','India'])],
x='date', y='daily_vaccinations_per_million',␣
↪hue='country')
chart.xaxis.set_major_locator(ticker.LinearLocator(10)) #adjust x-axis label to␣

↪make it readable
chart.ticklabel_format(style='plain', axis='y') #change scientific notation on␣

chart.set_xlabel("Date", fontsize = 15) #label x-axis

chart.set_ylabel("Daily Vaccination", fontsize = 15) #label y-axis
25
chart.set_title("Daily Vaccination Trends over Time (per million)", fontsize =␣
↪22) #set chart title
[ ]: Text(0.5, 1.0, 'Daily Vaccination Trends over Time (per million)')
The graph above shows that top countries in the GDP ranking (United States,China, Japan, Ger-
many, United Kingdom, India) has a higher growth rate in daily vaccination from the initial point.
However, the bottom countries in the GDP ranking (Venezuela, Rwanda, Paraguay,Cambodia) are
delayed in their growth rate in daily vaccinations about 6 months after the top countries in the
GDP ranking.
0.8 GeoGraph Preparation

[ ]: # Install and import the package
!pip install plotly_express -q
import plotly_express as px #interactive graph that oyou can hover over
[ ]: # Downsizing the table for relevant columns

country_vaccine_ffill = country_vaccine[['country', 'iso_code',␣
↪'people_fully_vaccinated_per_hundred', 'date']]
country_vaccine_ffill = country_vaccine_ffill.sort_values(by = 'date',␣

↪ascending = True)
# Creating helper variables for the for loop

country_name_list = country_vaccine_ffill.country.unique()
df_list = []
# Main loop to forward fill people_fully_vaccinated_per_hundred by country

for country in country_name_list:
temp_table = country_vaccine_ffill[country_vaccine_ffill.country == country]
# Making sure that we don't forward fill NaN but instead we forward fill 0 if␣
↪the first value is NaN
26
if pd.isna(temp_table.people_fully_vaccinated_per_hundred.iloc[0]):
temp_table.people_fully_vaccinated_per_hundred.iloc[0] = 0
temp_table = temp_table.ffill(axis=0)
df_list.append(temp_table)
country_vaccine_ffill = pd.concat(df_list)
/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py:1732:
SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-

docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
self._setitem_single_block(indexer, value, name)
[ ]: # Convert 'date' column from object type to datetime64 type

country_vaccine_ffill["date"] = pd.to_datetime(country_vaccine_ffill["date"])
[ ]: # Filter out dates from 2022-02-23 onwards

filtered_country_vaccine_ffill =␣
↪country_vaccine_ffill[country_vaccine_ffill['date'] < '2022-02-23']
[ ]: # Convert 'date' column from datetime64 type to object type

filtered_country_vaccine_ffill['date'] = filtered_country_vaccine_ffill['date'].
↪dt.strftime('%Y-%m-%d');
/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:2:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

0.8.1 Geographing Fully Vaccinated Populations Overtime
[ ]: fig = px.choropleth(filtered_country_vaccine_ffill, locations="iso_code",␣

↪color="people_fully_vaccinated_per_hundred",
hover_name="country", animation_frame="date",␣
↪color_continuous_scale='Plasma_r', range_color=[0,130])
fig.show()
0.9 Exploratory Data Analysis Summary

As we see overtime, the first countries providing vaccines are Russia, USA, Canada and Asian
countries, where 10% of their citizens received a jab and fully vaccinated. 6 months later, we can
27
see on the map that the rest of the world starts giving vaccinations, resulting in 10 per hundred
fully vaccinations citizens in Africa, 10% in Europe, 40% in Russia and 50% in the USA. 1 year
later, Canada, Australia and China are best in class and Africa is still lacking behind. The same is
true for 15 months later since start.
By the end of 2021, despite North America having the headstart for fully vaccinated
people, the EU quickly responds and surpasses from a global standpoint. After some
extensive research regarding their rollout plans, it appears that the EU invested in
the acceleration of vaccine companies to secure vaccine availability for their countrymen
(https://ec.europa.eu/commission/presscorner/detail/en/fs_20_1911). Having multiple vaccines
available to everyone allowed their union to get ahead on becoming a fully vaccinated population.
Africa does not appear to have a vigilant response towards COVID vaccines and is stated in having:
“poor health infrastructure, a lack of funding for training and deploying medical staff, as well as
vaccine storage issues” as contributing factors. (https://www.bbc.com/news/56100076)
Much of the decline in vaccination rates have been reported under the theory of vaccine hesitancy
given the severity of the pandemic, number of discovered COVID strains, as well as the emergency
use of vaccines that were pending FDA approval.
1 Next Steps
Our goal is to track COVID-19 vaccine adoption in the World. We will further analyze the rate of
people getting vaccinated per day in each country and the types of vaccine used by incorporating the
financial impacts. We planned to include pricing and vaccine availability (on a regression model)
to further predict the progress of vaccination and immunization status in these countries. In terms
of cleansing the data, we plan to filter out the zero’s from the dataset.
2 Data Modelling
2.0.1 Additional Datasets
We were able to find accurate Pricing trends from the COVID-19 Vaccine Market Dashboard.
However, due to inconsistency in the formatting and metrics available in the initial datasets, we
were unable to include pricing trends and measure its impact on vaccination and immunization
status.
[ ]: #vaccine_pricing = pd.read_csv("drive/MyDrive/Business Analytics_IS833_Team␣
↪Project/Vaccine_Pricing.csv")
#vaccine_pricing.head()
[ ]: #vaccine_pricing=vaccine_pricing.groupby(['vaccine'])[['Price per dose']].mean()

#vaccine_pricing.reset_index(inplace=True)
#vaccine_pricing
2.0.2 Data Preparation for Modelling

1. Before applying any machine learning model, we did a quick check to get a sense of the
predictive power of our features. Hence, we used the backward elimination method, to remove
28
the features that were not statistically significant. Upon using this method, we found that
all variables were statistically significant, and hence did not have to drop any feature.
2. Then we split the data into independant and dependant variables, with our dependant or
target variable being vax_predict, which is based on column fully vaccinated people per 100.
3. Next we split the data into train and test with a 70-30 split, and apply different models to it.
4. Logistic Regression- The accuracy for Logistic Rgression model is the highest which is 99.4%;
TPR (Specificity) is 99.92%; TNR sensitivity is 99.95%.
5. Further, we apply the GaussianNB model and get an accuracy score of 96.87%, Specificity of
98.71% and Sensitivity of 95.47%
6. Overall, we see that Logistic Regression is the most suited model and provides the best results.
[ ]: # Downsizing the table for relevant columns
#country_vaccine_ffill = country_vaccine[['country', 'iso_code',␣
↪'people_fully_vaccinated_per_hundred', 'date']]
country_vaccine_ffill = country_vaccine[['country', 'iso_code', 'date',

'total_vaccinations_per_hundred', 'people_vaccinated_per_hundred',
'people_fully_vaccinated_per_hundred', 'daily_vaccinations_per_million']]
country_vaccine_ffill = country_vaccine_ffill.sort_values(by = 'date',␣
↪ascending = True)
# Creating helper variables for the for loop

country_name_list = country_vaccine_ffill.country.unique()
df_list = []
# Main loop to forward fill people_fully_vaccinated_per_hundred by country

for country in country_name_list:
temp_table = country_vaccine_ffill[country_vaccine_ffill.country == country]
# Making sure that we don't forward fill NaN but instead we forward fill 0 if␣
↪the first value is NaN
if pd.isna(temp_table.people_fully_vaccinated_per_hundred.iloc[0]):
temp_table.people_fully_vaccinated_per_hundred.iloc[0] = 0
temp_table = temp_table.ffill(axis=0)
df_list.append(temp_table)
country_vaccine_ffill = pd.concat(df_list)
country_vaccine_ffill.head()
/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py:1732:
A value is trying to be set on a copy of a slice from a DataFrame

29
[ ]: country iso_code date total_vaccinations_per_hundred \
58517 Norway NOR 2020-12-02 0.0
58518 Norway NOR 2020-12-03 0.0
58519 Norway NOR 2020-12-04 0.0
58520 Norway NOR 2020-12-05 0.0
58521 Norway NOR 2020-12-06 0.0
people_vaccinated_per_hundred people_fully_vaccinated_per_hundred \
58517 0.0 0.0
58518 0.0 0.0
58519 0.0 0.0
58520 0.0 0.0
58521 0.0 0.0
daily_vaccinations_per_million
58517 NaN
58518 0.0
58519 0.0
58520 0.0
58521 0.0
[ ]: country_vaccine_ffill.info()
Int64Index: 86512 entries, 58517 to 12644
--- ------ -------------- -----
0 country 86512 non-null object
1 iso_code 86512 non-null object
dtypes: float64(4), object(3)
[ ]: cols =␣
↪['total_vaccinations_per_hundred','people_vaccinated_per_hundred','people_fully_vaccinated_p
country_vaccine_ffill = country_vaccine_ffill.sort_values(cols).groupby(cols,␣
↪as_index=False).max()
[ ]: #country_vaccine_ffill.
↪loc[(country_vaccine_ffill['people_fully_vaccinated_per_hundred'] <= 50) &␣
↪(country_vaccine_ffill['people_fully_vaccinated_per_hundred'] >=10),␣
↪'fully_vaccinated_groups'] = 1
30
country_vaccine_ffill.
↪loc[country_vaccine_ffill['people_fully_vaccinated_per_hundred'] <=␣
↪country_vaccine_ffill['people_fully_vaccinated_per_hundred'].mean(),␣
country_vaccine_ffill.
↪loc[country_vaccine_ffill['people_fully_vaccinated_per_hundred'] >␣
↪country_vaccine_ffill['people_fully_vaccinated_per_hundred'].mean() ,␣
country_vaccine_ffill.head()
[ ]: total_vaccinations_per_hundred people_vaccinated_per_hundred \
0 0.0 0.0
1 0.0 0.0
2 0.0 0.0
3 0.0 0.0
4 0.0 0.0
0 0.0 0.0
1 0.0 1.0
2 0.0 2.0
3 0.0 3.0
4 0.0 4.0
country iso_code date fully_vaccinated_groups

0 Switzerland NZL 2021-02-18 0.0
1 Venezuela VEN 2021-02-22 0.0
2 Thailand THA 2021-04-13 0.0
3 Vietnam VNM 2021-04-28 0.0
4 Vietnam VNM 2021-10-28 0.0
[ ]: df=country_vaccine_ffill.drop(['country','iso_code','date'],axis=1)
#df['date'] = pd.to_datetime(df['date'])
df.fillna(0,inplace=True)
df.info()
df.isnull().sum()
--- ------ -------------- -----
4 fully_vaccinated_groups 70279 non-null float64
31
dtypes: float64(5)
memory usage: 2.7 MB
[ ]: total_vaccinations_per_hundred 0
people_vaccinated_per_hundred 0
people_fully_vaccinated_per_hundred 0
daily_vaccinations_per_million 0
fully_vaccinated_groups 0
dtype: int64
[ ]: df.rename(columns = {'total_vaccinations_per_hundred':'t_vax_rate',␣
↪'people_vaccinated_per_hundred':'people_vax',
'people_fully_vaccinated_per_hundred':
↪'people_2_vax','daily_vaccinations_per_million':
↪'daily_rate','fully_vaccinated_groups':'vax_predict'}, inplace = True)
[ ]: #df['date']=df['date'].astype('int')
[ ]: import pandas as pd
writer = df.to_excel("output.xlsx",
sheet_name='Sheet_name_1',index=False)
We tried to include the time period in the model, however, the data quality is still not consistent
for different countries over specific durations. With such limitations, we decided to incorporate the
features only for the latest available date. We observed (while running regression analysis in Excel)
that all the independent variables are statistically significant, such that, p <= 5% and T stat >=2.
2.0.3 Correlation Coefficients and Features Matrix

[ ]: corrMatrix = round(df.corr(),2)
sns.set(rc={'figure.figsize':(14,9)})
corr_heatmap=sns.heatmap(corrMatrix, annot=True,cmap="Blues")
corr_heatmap.set_title('Correlation Coefficients');
plt.show()
32
[ ]: df.hist(figsize=(15,10),bins=10)
plt.show()
33
[ ]: import seaborn as sns
import matplotlib.pyplot as plt
sns.color_palette("Spectral", as_cmap=True)
sns.pairplot(df)
[ ]: <seaborn.axisgrid.PairGrid at 0x7f1e7a2db450>
34
Splitting Test & Train Dataset
[ ]: y = df['vax_predict']
X=df.drop('vax_predict', axis=1)
from sklearn.model_selection import train_test_split
Xtrain, Xtest, ytrain, ytest = train_test_split(X, y, test_size=0.3,␣
↪random_state=833)
Modelling using Logistic Regression and Checking accuracy

[ ]: from sklearn.linear_model import LogisticRegression
from sklearn.metrics import confusion_matrix,accuracy_score
model = LogisticRegression(solver='liblinear') # 2. instantiate model
35
model.fit(Xtrain, ytrain) # 3. fit model to data
y_model = model.predict(Xtest) # 4. predict on new data
accuracy_score(ytest, y_model)
[ ]: 0.9994308480364257
[ ]: sns.set(rc={'figure.figsize':(8,5)})
conf_matrix=confusion_matrix(ytest,y_model)
conf_heatmap=sns.heatmap(conf_matrix, annot=True,fmt='d',cmap="OrRd")
conf_heatmap.set_title('Confusion Matrix');
plt.show()
[ ]: P = sum(ytest == 1)
TP = sum((ytest == 1) & (y_model == 1))
TPR =TP/P
TPR
[ ]: 0.9992318665642489
[ ]: N = sum(ytest == 0)
TN = sum((ytest == 0) & (y_model == 0))
TNR=TN/N
TNR
36
[ ]: 0.9995823239495447
Modelling using Gaussian Naive Bayes and Checking accuracy

[ ]: from sklearn.naive_bayes import GaussianNB
model = GaussianNB() # 2. instantiate model
[ ]: 0.9687440713337128
[ ]: conf_matrix=confusion_matrix(ytest,y_model)
plt.show()
[ ]: P = sum(ytest == 1)
TPR =TP/P
TPR
37
[ ]: 0.9871611982881597
[ ]: N = sum(ytest == 0)
TNR=TN/N
TNR
[ ]: 0.9547239161306491
Modelling using Decision Tree Classifier and Checking accuracy

[ ]: from sklearn.tree import DecisionTreeClassifier

model = DecisionTreeClassifier() # 2. instantiate model
[ ]: 1.0
[ ]: conf_matrix=confusion_matrix(ytest,y_model)
plt.show()
38
[ ]: P = sum(ytest == 1)
TPR =TP/P
TPR
[ ]: 1.0
[ ]: N = sum(ytest == 0)
TNR=TN/N
TNR
[ ]: 1.0
3 Recommendation
Through the analysis of this dataset, we have the following recommendations:
• In order to make better financial estimations for vaccination strategies, we recommend in-
cluding data for features, such as, pricing, dosage frequency, dosage amount (ml), logistic
limitations, etc.
• Based on further analysis, we observed that the data for African countries is available and we
recommend including the details from other reliable data sources. African countries have the
vaccination rate of below 10% threshold but was not reflected in the dataset we analyzed.(ref).
With improved dataset, the model would recommend increase in the vaccination rate in
African countries.
Increase the vaccination rate in Uruguay, Bulgaria, Slovenia, Latvia, Estonia, Cyprus, Malta,
Luxembourg, Iceland, Liechtenstein. These are countries with lowest vaccination rates.
We should encourage countries to keep vaccinating citizens with at least 2 doses of vaccine (3 doses
encouraged) in order to fight against any new variant of Corona virus. Big brands like Pfizer,
Moderna, Astrazeneca should produce more vaccine and distribute non-profitably to countries (like
African countries).
4 Dashboard
COVID-19 Vaccinations Dashboard - Google Studio
39

Rogress of Covid 19 Vaccination 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rogress of Covid 19 Vaccination 1

Uploaded by

Copyright:

Available Formats

rogress-of-covid-19-vaccination-1

April 20, 2024

#Analyzing and Predicting Progress of COVID-19 Vaccination

IS833 Business Analytics Project - Team 5

0.1 Dataset link:

Original Source: COVID-19 Dataset by Our World in Data (link)

0.2 Introduction & Project Objective

0.4 Data importation

# Enlarge charts produced

[ ]: from google.colab import drive

[ ]: # Upload file from Kaggle to be read here

country_vaccine = pd.read_csv("drive/MyDrive/Business Analytics_IS833_Team␣

company_vaccine = pd.read_csv("drive/MyDrive/Business Analytics_IS833_Team␣

[ ]: country iso_code date total_vaccinations people_vaccinated \

people_fully_vaccinated daily_vaccinations_raw daily_vaccinations \

[ ]: location date vaccine total_vaccinations

0.5 Data Categories

0.6 Data Cleaning and Validation Checks

# Display country_vaccine with new date format

[ ]: country iso_code date total_vaccinations people_vaccinated \

people_fully_vaccinated daily_vaccinations_raw daily_vaccinations \

[ ]: ## Sample code to reformat DMY to YMD (^Now changed to pandas function)

## Reformat date from DMY to YMD

[ ]: # Display maximum date in timeline

[ ]: # Display minimum date in timeline

[ ]: vaccine CanSino Covaxin Johnson&Johnson Moderna Novavax \

vaccine Oxford/AstraZeneca Pfizer/BioNTech Sinopharm/Beijing \

vaccine Sinovac Sputnik V

[ ]: # India data is not available in the 2nd table␣

[ ]: #Display the latest values for each Country-Vaccine combination

[ ]: location country_max vaccine total_vaccinations

[ ]: location country_max vaccine total_vaccinations Rank

0.7 Exploratory Data Analysis

barplot = sns.barplot(x='location', y='total_vaccinations',␣

barplot.ticklabel_format(style='plain', axis='y') #change scientific notation␣

barplot.tick_params(axis='x', rotation=90) #NV changed to not have overlapping␣

barplot.set_ylabel("Total Vaccinations", fontsize = 12) #label x-axis

#plt.xlim(0, 650000000) #set same x-axis across other vaccines

# Plot total vaccinations per country using specific vaccine

barplot = sns.barplot(x='location', y='total_vaccinations',␣

barplot.ticklabel_format(style='plain', axis='y') #change scientific notation␣

barplot.set_ylabel("Total Vaccinations", fontsize = 12) #label x-axis

#plt.xlim(0, 650000000) #set same x-axis across other vaccines

df.sort_values(by='total_vaccinations', ascending=False, inplace=True)

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

df.sort_values(by='total_vaccinations', ascending=False, inplace=True)

barplot = sns.barplot(y=df_top10.index, x='total_vaccinations', data=df_top10)

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

df.sort_values(by='total_vaccinations', ascending=False, inplace=True)

barplot = sns.barplot(y=df_top10.index, x='total_vaccinations', data=df_top10)

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

barplot.set_title("Countries with highest usage of Oxford/AstraZeneca Vaccine",␣

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣

barplot.set_xlabel("Total Vaccinations", fontsize = 15) #label x-axis

df.sort_values(by='total_vaccinations', ascending=False, inplace=True)

barplot = sns.barplot(y=df_top10.index, x='total_vaccinations', data=df_top10)

barplot.ticklabel_format(style='plain', axis='x') #change scientific notation␣