Covid-19 Analysis and Visualization using Plotly Express

In this article, we will discuss Analyse Covid-19 data and will visualize it using Plotly Express in Python. This article deals with creating dozens of bar charts, line graphs, bubble charts, scatter plots. The graph that will be made in this project will be of excellent quality. Envisioning COVID-19 will primarily be using Plotly Express for this project. The analysis and visualization enable people to understand complex scenarios and make predictions about the future from the current situation.

This analysis summarizes the modeling, simulation, and analytics work around the COVID-19 outbreak around the world from the perspective of data science and visual analytics. It examines the impact of best practices and preventive measures in various sectors and enables outbreaks to be managed with available health resources.

Tools and Technologies Used in the Project: Google Colab(Runtime type - GPU).

Requirements to Build the Project: 

• Basic knowledge of Python
• Basic understanding of graphs and charts
• Data visualization
• Pandas
• Numpy
• Matplotlib
• Plotly Express
• Choropleth
• Wordcloud

Stepwise Implementation

Step 1: Importing Necessary Libraries

The task is simple, once the installation of all the required libraries is successful, they need to be imported to the working space, since they will provide the additional support for analysis and visualization.

Example: importing libraries

# Data analysis and Manipulation
import plotly.graph_objs as go
import plotly.io as pio
import plotly.express as px
import pandas as pd

# Data Visualization
import matplotlib.pyplot as plt

# Importing Plotly
import plotly.offline as py
py.init_notebook_mode(connected=True)

# Initializing Plotly
pio.renderers.default = 'colab'

Step 2: Importing the Datasets

Importing three datasets into this project

• covid- This dataset contains Country/Region, Continent,  Population, TotalCases, NewCases, TotalDeaths, NewDeaths,  TotalRecovered, NewRecovered, ActiveCases, Serious, Critical, Tot Cases/1M pop, Deaths/1M pop, TotalTests, Tests/1M pop, WHO Region, iso_alpha.
• covid_grouped- This dataset contains Date(from 20-01-22 to 20-07-27), Country/Region, Confirmed, Deaths, Recovered, Active, New cases, New deaths, New recovered, WHO Region, iso_alpha.
• coviddeath- This dataset contains real-world examples of a number of Covid-19 deaths and the reasons behind the deaths.

To import datasets to the working space pandas read_csv() method can be used.

Syntax:

read_csv(path)

Example: Importing datasets

# Importing Dataset1
dataset1 = pd.read_csv("covid.csv")
dataset1.head()  # returns first 5 rows

Output:

Further, information regarding the dataset we are using will help us sample it better for analysis.

Example: getting dataset information

# Returns tuple of shape (Rows, columns)
print(dataset1.shape)

# Returns size of dataframe
print(dataset1.size)

Output:

(209, 17)
3553

Example: dataset information

# Information about Dataset1
# return concise summary of dataframe
dataset1.info()  

Output:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 209 entries, 0 to 208
Data columns (total 17 columns):
 #   Column            Non-Null Count  Dtype  
---  ------            --------------  -----  
 0   Country/Region    209 non-null    object 
 1   Continent         208 non-null    object 
 2   Population        208 non-null    float64
 3   TotalCases        209 non-null    int64  
 4   NewCases          4 non-null      float64
 5   TotalDeaths       188 non-null    float64
 6   NewDeaths         3 non-null      float64
 7   TotalRecovered    205 non-null    float64
 8   NewRecovered      3 non-null      float64
 9   ActiveCases       205 non-null    float64
 10  Serious,Critical  122 non-null    float64
 11  Tot Cases/1M pop  208 non-null    float64
 12  Deaths/1M pop     187 non-null    float64
 13  TotalTests        191 non-null    float64
 14  Tests/1M pop      191 non-null    float64
 15  WHO Region        184 non-null    object 
 16  iso_alpha         209 non-null    object 
dtypes: float64(12), int64(1), object(4)
memory usage: 27.9+ KB

Similarly other datasets can be imported and explored.

Example: Importing dataset 

# Importing Dataset2
dataset2 = pd.read_csv("covid_grouped.csv")
dataset2.head()  # return first 5 rows of dataset2

Output:

Example: Getting dataset information

# Returns tuple of shape (Rows, columns)
print(dataset2.shape)

# Returns size of dataframe
print(dataset2.size)

Output:

(35156, 11)
386716

Example: Dataset information

# Information about Dataset2
dataset2.info()  # return concise summary of dataframe

Output:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 35156 entries, 0 to 35155
Data columns (total 11 columns):
 #   Column          Non-Null Count  Dtype 
---  ------          --------------  ----- 
 0   Date            35156 non-null  object
 1   Country/Region  35156 non-null  object
 2   Confirmed       35156 non-null  int64 
 3   Deaths          35156 non-null  int64 
 4   Recovered       35156 non-null  int64 
 5   Active          35156 non-null  int64 
 6   New cases       35156 non-null  int64 
 7   New deaths      35156 non-null  int64 
 8   New recovered   35156 non-null  int64 
 9   WHO Region      35156 non-null  object
 10  iso_alpha       35156 non-null  object
dtypes: int64(7), object(4)
memory usage: 3.0+ MB

Step 3: Dataset cleaning

Data cleaning is the process of altering, modifying a recordset, correcting erroneous records from the database and identifying incomplete, incorrect, or irrelevant parts of the data, and then removing dirty data.

Example: Getting dataset columns

# Columns labels of a Dataset1
dataset1.columns

Output:

Index(['Country/Region', 'Continent', 'Population', 'TotalCases', 'NewCases', 'TotalDeaths', 'NewDeaths', 'TotalRecovered', 'NewRecovered',  'ActiveCases', 'Serious,Critical', 'Tot Cases/1M pop', 'Deaths/1M pop', 'TotalTests', 'Tests/1M pop', 'WHO Region', 'iso_alpha'], dtype='object')

We don't need 'NewCases', 'NewDeaths', 'NewRecovered' columns as they contains NaN values. So drop these columns by drop() function of pandas.

Syntax:

drop(name of columns)

Example: Cleaning dataframe

# Drop NewCases, NewDeaths, NewRecovered rows from dataset1

dataset1.drop(['NewCases', 'NewDeaths', 'NewRecovered'], 
              axis=1, inplace=True)

# Select random set of values from dataset1
dataset1.sample(5)

Output:

Let's create a table through the table function already available in plotly express.

Example: Creating table using plotly express

# Import create_table Figure Factory

from plotly.figure_factory import create_table

colorscale = [[0, '#4d004c'], [.5, '#f2e5ff'], [1, '#ffffff']]
table = create_table(dataset1.head(15), colorscale=colorscale)
py.iplot(table)

Output:

Step 4: Bar graphs- Comparisons between COVID infected countries in terms of total cases, total deaths, total recovered & total tests

Using one line of code, we will create amazing graphs using Plotly Express. Visualization can be done easily by moving the cursor in any plot, we can get label presence point directly by using the cursor. We can visualize and analyze the dataset with each aspect using the relation between the columns.

Primarily look at the country with respect to a total number of cases by top 15 countries only and color total cases and hover data as 'Country/Region', 'Continent'.

Example: Bar graph 

px.bar(dataset1.head(15), x = 'Country/Region', 
       y = 'TotalCases',color = 'TotalCases', 
       height = 500,hover_data = ['Country/Region', 'Continent'])

Output:

As the plot clearly shows the data for the top 15 countries, now again take the country with respect to the total number of cases from the top 15 countries, color the total deaths hover data as 'Country/Region', 'Continent' and analyze the visualization.

Example: Bar graph

px.bar(dataset1.head(15), x = 'Country/Region', y = 'TotalCases',
       color = 'TotalDeaths', height = 500,
       hover_data = ['Country/Region', 'Continent'])

Output:

Let's analyze by coloring the total number of recovered cases 

Example: Bar graph

px.bar(dataset1.head(15), x = 'Country/Region', y = 'TotalCases',
       color = 'TotalDeaths', height = 500,
       hover_data = ['Country/Region', 'Continent'])

Output:

Visualize the same again by coloring the total number of tests.

Example: Bar graph

px.bar(dataset1.head(15), x = 'Country/Region', y = 'TotalCases',
       color = 'TotalTests', height = 500, hover_data = ['Country/Region', 'Continent'])

Output:

The visualization could be as we have done with the top 15 countries with total cases, deaths, recoveries, and tests. We can analyze the plot by looking at them.

Let's create a horizontal orientation plot with X-axis as 'TotalTests' and Y-axis as 'Country/Region' with passing parameter orientation="h" and color the plot by 'TotalTests'.

Example: Bar graph 

px.bar(dataset1.head(15), x = 'TotalTests', y = 'Country/Region',
       color = 'TotalTests',orientation ='h',  height = 500,
       hover_data = ['Country/Region', 'Continent'])

Output:

Let's look at 'TotalTests' followed by 'Continent' and color the plot with 'Continent'.

Example: Bar graph

px.bar(dataset1.head(15), x = 'TotalTests', y = 'Continent',
       color = 'TotalTests',orientation ='h',  height = 500,
       hover_data = ['Country/Region', 'Continent'])

Output:

Step 5: Data Visualization through Bubble Charts-Continent Wise

Let's create a scatter plot and take a look at the continent's statistics, firstly look at the total number of cases by continent and take hover data as 'Country/Region', 'Continent'.

Example: Scatter plot

px.scatter(dataset1, x='Continent',y='TotalCases', 
           hover_data=['Country/Region', 'Continent'], 
           color='TotalCases', size='TotalCases', size_max=80)

Output:

log_y= True, the histogram axis (not the returned parameter) is in log scale. The return parameter (n, bins), i.e. the values of bins and sides of bins are the same for log=True and log=False. This means both n==n2 and bins==bins2 are true

Example: Scatter plot

px.scatter(dataset1.head(57), x='Continent',y='TotalCases', 
           hover_data=['Country/Region', 'Continent'], 
           color='TotalCases', size='TotalCases', size_max=80, log_y=True)

Output:

Example: Scatter plot

px.scatter(dataset1.head(54), x='Continent',y='TotalTests', 
           hover_data=['Country/Region', 'Continent'], 
           color='TotalTests', size='TotalTests', size_max=80)

Output:

Example: Scatter plot

px.scatter(dataset1.head(50), x='Continent',y='TotalTests', 
           hover_data=['Country/Region', 'Continent'], 
           color='TotalTests', size='TotalTests', size_max=80, log_y=True)

Output:

Step 6: Data Visualization through Bubble Charts-Country Wise

Let's take a look at the country-wise data visualization, first look at the continent with respect to the total number of deaths by the top 50 countries only and color the total number of deaths and take the hover data as 'Country/Region', 'Continent'.

Example: Bubble chart

px.scatter(dataset1.head(100), x='Country/Region', y='TotalCases', 
           hover_data=['Country/Region', 'Continent'],
           color='TotalCases', size='TotalCases', size_max=80)

Output:

Now, the Country/Region with respect to the total number of cases for top 30 countries only and color the total number of cases and take the hover data as 'Country/Region', 'Continent'.

Example: Bubble chart

px.scatter(dataset1.head(30), x='Country/Region', y='TotalCases', 
           hover_data=['Country/Region', 'Continent'],
           color='Country/Region', size='TotalCases', size_max=80, log_y=True)

Output:

Now format the image of the country/region in relation to the total number of deaths. And do the same for the other aspects of COVID-19 from dataset1.

Example: Bubble chart

px.scatter(dataset1.head(10), x='Country/Region', y= 'TotalDeaths', 
           hover_data=['Country/Region', 'Continent'],
           color='Country/Region', size= 'TotalDeaths', size_max=80)

Output:

Example: Bubble chart

px.scatter(dataset1.head(30), x='Country/Region', y= 'Tests/1M pop', 
           hover_data=['Country/Region', 'Continent'],
           color='Country/Region', size= 'Tests/1M pop', size_max=80)

Output:

Example: Country/Region VS Tests/1M pop (color-scale of Tests/1M pop)

px.scatter(dataset1.head(30), x='Country/Region', y= 'Tests/1M pop', 
           hover_data=['Country/Region', 'Continent'],
           color='Tests/1M pop', size= 'Tests/1M pop', size_max=80)

Output: 

Example: Bubble chart

px.scatter(dataset1.head(30), x='TotalCases', y= 'TotalDeaths', 
           hover_data=['Country/Region', 'Continent'],
           color='TotalDeaths', size= 'TotalDeaths', size_max=80)

Output:

It is clear from the result that they have a linear relationship between the total number of cases and the total number of deaths. That means more cases, more deaths.

Example: Bubble chart

px.scatter(dataset1.head(30), x='TotalCases', y= 'TotalDeaths', 
           hover_data=['Country/Region', 'Continent'],
           color='TotalDeaths', size= 'TotalDeaths', size_max=80, 
           log_x=True, log_y=True)

Output:

Example: Bubble chart

px.scatter(dataset1.head(30), x='TotalTests', y= 'TotalCases', 
           hover_data=['Country/Region', 'Continent'],
           color='TotalTests', size= 'TotalTests', size_max=80, 
           log_x=True, log_y=True)

Output:

Step 7: Advanced Data Visualization- Bar graphs for All top infected Countries

In this task, we will explore covid-19 data using bar graphs and charts and use dataset2 as it has date column.

Example: Bar chart

px.bar(dataset2, x="Date", y="Confirmed", color="Confirmed", 
       hover_data=["Confirmed", "Date", "Country/Region"], height=400)

Output:

The above graph we get as output which includes all countries with respect to recovered cases. we can imagine the exponential growth of corona cases by date. We can use log function for this to be more clear. 

Example: Bar chart

px.bar(dataset2, x="Date", y="Confirmed", color="Confirmed", 
       hover_data=["Confirmed", "Date", "Country/Region"],log_y=True, height=400)

Output:

Let's imagine death instead of confirmation with the same and color it by date.

Example: Bar chart

px.bar(dataset2, x="Date", y="Deaths", color="Deaths", 
       hover_data=["Confirmed", "Date", "Country/Region"],
       log_y=False, height=400)

Output:

Step 8: Countries Specific COVID Data Visualization: (United States)

In this specific task, we will analyze data of the USA country.

Example: Refining dataset to get only USA data 

df_US= dataset2.loc[dataset2["Country/Region"]=="US"]

Now let us plot and study the covid situation in the USA.

Example: Bar chart

px.bar(df_US, x="Date", y="Confirmed", color="Confirmed", height=400)

Output:

Here we can clearly see how the confirmed cases increased in the United States with respect to time (January 2020 to July 2020). Similarly, we can check the same for recovered cases, tests and deaths.

Example: Bar chart

px.bar(df_US,x="Date", y="Recovered", color="Recovered", height=400)

Output:

Similarly, we can analyze the data in all the ways to generate the line graph for the same.

Example: Line plot

px.line(df_US,x="Date", y="Recovered", height=400)

Output:

Example: Line plot

px.line(df_US,x="Date", y="Deaths", height=400)

Output:

Example: Line plot

px.line(df_US,x="Date", y="Confirmed", height=400)

Output:

Example: Line plot

px.line(df_US,x="Date", y="New cases", height=400)

Output:

Now let us create bar charts and study the USA situation using that.

Example: Bar chart

px.bar(df_US,x="Date", y="New cases", height=400)

Output:

Similarly, let us also plot a line plot.

Example: Scatter plot

px.scatter(df_US, x="Confirmed", y="Deaths", height=400)

Output:

Step 9: Visualization of Data in terms of Maps

We can use choropleth to visualize the data in terms of maps, with maps usually being the predominant way of visualizing the data. Since COVID-19 is a global phenomenon and so we look through and fix them in terms of wall maps. Ortho-graphics, rectangular and natural earth projection to visualize the data With dataset2 for the purpose as it has Dates column. It will look at the growth of Covid-19 (from Jan to July 2020) as in how the virus reached across the world.

Choropleth is an amazing representation of data on a map. Choropleth maps provide an easy way to visualize how a measurement varies across a geographic areal-Life

Project Application in Real choropleth map displays divided geographical areas or regions that are colored, shaded or patterned in relation to a data variable.

Equi-rectangular Projection:

Syntax: chloropleth()

parameters:

• dataset
• locations= ISOALPHA
• color
• hover_name
• color_continuous_scale= [RdYlGn, Blues, Viridis...]
• animation_frame= Date

Example: Creating map

px.choropleth(dataset2,
              locations="iso_alpha",
              color="Confirmed",
              hover_name="Country/Region", 
              color_continuous_scale="Blues",
              animation_frame="Date")

Output:

This creates an animation containing visualizations from January to July 2020. Playing this animation will make it more clear how the virus spread around the world. The darker the color, the higher the confirmed cases are.

Example: Creating map

px.choropleth(dataset2,
              locations='iso_alpha',
              color="Deaths",
              hover_name="Country/Region",
              color_continuous_scale="Viridis",
              animation_frame="Date" )

Output:

This code creates an animation of death cases by date. By playing this animation it will be shown how deaths increase around the world.

Natural Earth projection is a compromise pseudo-cylindrical map projection for world maps.

Example: Natural earth projection

px.choropleth(dataset2,
              locations='iso_alpha',
              color="Recovered",
              hover_name="Country/Region",
              color_continuous_scale="RdYlGn",
              projection="natural earth",
              animation_frame="Date" )

Output:

By running the output, things start to become more clear about how the recovery rate changes with respect to the date. Lets also look at how an animation can be added to a bar graph. We can convert the bar graph into animation using Dates column that is in dataset2.

Example: Bar graph animation

px.bar(dataset2, x="WHO Region", y="Confirmed", color="WHO Region", 
       animation_frame="Date", hover_name="Country/Region")

Output:

When running the output, the animation will run from January to July 2020. It will show 6 different bar graphs, each continent has its own color representing the confirmed cases.

Step 10: Visualize text using Word Cloud

Visualize the causes of death due to covid-19, as covid-19 affects people in different ways, hence creating a word cloud to visualize the leading cause of covid-19 deaths. To visualize the text the steps need to be followed are-

• Used to convert data elements of an array into list.
• Convert the string to one single string.
• Convert the string into word cloud

Dataset3: This dataset contains real world examples of number of Covid-19 deaths and the reasons behind the deaths.

Example: Importing dataset

dataset3= pd.read_csv("coviddeath.csv")
dataset3.head()

Output:

Example: Dataset information

dataset3.tail()

Output:

Example: Getting dataset information

dataset3.groupby(["Condition"]).count()

Output:

Example: Creating wordcloud

# import word cloud
from wordcloud import WordCloud

sentences = dataset3["Condition"].tolist()
sentences_as_a_string = ' '.join(sentences)


# Convert the string into WordCloud
plt.figure(figsize=(20, 20))
plt.imshow(WordCloud().generate(sentences_as_a_string))

Output:

From the output, it can be clearly seen that the leading cause of death is Influenza Pneumonia. We have converted the  condition group to the list and stored the list in the variable  "column_to_list". Here we have converted the list into a single string and stored in a variable named "column2_to_string" by using .join().

Example: Creating wordcloud

column2_tolist= dataset3["Condition Group"].tolist()

# Convert the list to one single string
column_to_string= " ".join(column2_tolist)

# Convert the string into WordCloud
plt.figure(figsize=(20,20))
plt.imshow(WordCloud().generate(column_to_string))

Output:

Here, respiratory diseases are the major cause of death followed by circulatory diseases which are cardiovascular diseases.

Colab Link

To get the notebook, click here. For dataset, click here.

tanushree_dev

Follow

Improve

Article Tags :

• Data Science
• Machine Learning
• Blogathon
• AI-ML-DS
• Blogathon-2021
• python
• Data Visualization
• Python-Plotly

+4 More

Practice Tags :

• Machine Learning
• python