plotly on Software Development Notes

Countries with highest deaths from covid

Eric Callanan — Saturday, 18 Dec 2021

A lot of time was spent in 2020 looking at daily covid numbers, which was getting overwhelming. Having taken a break for a while, this article combines covid data and vaccination data with population data to present a view of the countries with the highest deaths from Covid-19 and the vaccination status in the country.

The Jupyter notebook is available on Github Countries-with-highest-deaths-from-covid-19-COVID-19 and it can be launched in a browser using Binder with this link Binder - Countries with highest deaths from COVID-19.

Load Covid data

As shown previously in Countries with highest confirmed cases of Covid-19 with Plotly the data is available from Johns Hopkins University who have made the data available on GitHub. More information about COVID-19 and the coronavirus is available from Coronavirus disease (COVID-19) advice for the public.

The latest data on deaths from Covid-19 is loaded directly from github.

 1# Number of Death Cases - Global: time series data for deaths
 2deaths_path = 'https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv'
 3
 4deaths_17_df = pd.read_csv(deaths_path, usecols=['Country/Region', '12/17/21'])
 5deaths_17_df = deaths_17_df.groupby("Country/Region").sum()
 6deaths_17_df.rename_axis(None, axis=0, inplace=True)
 7deaths_17_df
 8
 9"""
10                     12/17/21
11Afghanistan              7332
12Albania                  3158
13Algeria                  6175
14Andorra                   134
15Angola                   1738
16Antigua and Barbuda       117
17Argentina              116892
18Armenia                  7885
19Australia                2142
20Austria                 13438
21"""

 1Top_10 = deaths_17_df.sort_values(by = deaths_17_df.columns[-1], ascending = False).head(10)
 2Top_10
 3
 4"""
 5                12/17/21
 6US                805823
 7Brazil            617395
 8India             476869
 9Mexico            297356
10Russia            289292
11Peru              202076
12United Kingdom    147509
13Indonesia         143986
14Italy             135421
15Iran              130992
16"""

 1bg_color = 'rgba(208,225,242,1.0)'
 2latest_date = pd.to_datetime(deaths_17_df.columns[0])
 3
 4fig = go.Figure(
 5    data = [go.Bar(x = list(Top_10.iloc[:, 0]),
 6                   y = list(Top_10.index),
 7                   hovertemplate = "%{y}: <br>Deaths: %{x:,}<extra></extra>",
 8                   orientation = "h")]
 9)
10
11fig.update_layout(
12    # Set default font
13        font = dict(
14        family = "Droid Sans",
15        size = 14
16    ),
17    # Set figure title
18    title = dict(
19        text = f"<b>Countries with highest deaths from Covid [{latest_date.strftime('%b %d, %Y')}]</b>",
20        xref = 'container',
21        yref = 'container',
22        x = 0.5,
23        y = 0.87,
24        xanchor = 'center',
25        yanchor = 'middle',
26        font = dict(family = 'Droid Sans', size = 24)
27    ),
28    # set the plot bacground color
29    plot_bgcolor = bg_color,
30    # set the hover background color
31    hoverlabel = dict(
32        bgcolor = 'rgba(75,152,201,0.2)',
33        font_size = 16
34    ),
35    paper_bgcolor = bg_color,
36)
37
38fig.update_traces(marker_color = 'rgb(75,152,201)')
39fig['layout']['yaxis']['autorange'] = "reversed"
40fig.show()

Countries with highest deaths from COVID-19

Load Vaccination data

Covid-19 vaccination data is also available publicly on Github at COVID-19 data.

 1# Number of Vaccines - Global: time series data for deaths
 2vaccines_path = "https://raw.githubusercontent.com/govex/COVID-19/master/data_tables/vaccine_data/global_data/time_series_covid19_vaccine_global.csv"
 3vaccines_df = pd.read_csv(vaccines_path)
 4
 5print(f"Shape of vaccines = {vaccines_df.shape}")
 6"""
 7Shape of vaccines = (129128, 8)
 8"""
 9
10print(vaccines_df.iloc[[0,1,2,-3,-2,-1], :])
11"""
12       Country_Region        Date  Doses_admin  People_partially_vaccinated   People_fully_vaccinated Report_Date_String    UID Province_State 
130              Canada  2020-12-14          5.0                          0.0                       0.0         2020-12-14  124.0            NaN 
141               World  2020-12-14          5.0                          0.0                       0.0         2020-12-14    NaN            NaN 
152              Canada  2020-12-15        723.0                          0.0                       0.0         2020-12-15  124.0            NaN 
16129125          Yemen  2021-12-17     786027.0                     556652.0                  366587.0         2021-12-18  887.0            NaN 
17129126         Zambia  2021-12-17    1342989.0                     806611.0                  904201.0         2021-12-18  894.0            NaN 
18129127       Zimbabwe  2021-12-17    7058474.0                    4023289.0                 3035185.0         2021-12-18  716.0            NaN 
19"""

Look at vaccinations for a single Country in the most recent dates.

 1vaccines_df[vaccines_df['Country_Region'] == 'Mexico'].tail(10)
 2"""
 3
 4       Country_Region        Date  Doses_admin  People_partially_vaccinated  People_fully_vaccinated Report_Date_String    UID Province_State 
 5123215         Mexico  2021-12-08  134370326.0                   78816099.0               65543077.0         2021-12-09  484.0            NaN 
 6123849         Mexico  2021-12-09  135022230.0                   78816099.0               65543077.0         2021-12-10  484.0            NaN 
 7124483         Mexico  2021-12-10  135481459.0                   78816099.0               65543077.0         2021-12-11  484.0            NaN 
 8125117         Mexico  2021-12-11  136829147.0                   80191983.0               66003384.0         2021-12-12  484.0            NaN 
 9125751         Mexico  2021-12-12  137169511.0                   80388496.0               66150375.0         2021-12-13  484.0            NaN 
10126385         Mexico  2021-12-13  137357032.0                   80503005.0               66225140.0         2021-12-14  484.0            NaN 
11127019         Mexico  2021-12-14  137357032.0                   80503005.0               66225140.0         2021-12-15  484.0            NaN 
12127651         Mexico  2021-12-15  137357032.0                   81770394.0               66459570.0         2021-12-16  484.0            NaN 
13128283         Mexico  2021-12-16  139820373.0                   81770394.0               66459570.0         2021-12-17  484.0            NaN 
14128915         Mexico  2021-12-17  139820373.0                   81785946.0               66586509.0         2021-12-18  484.0            NaN 
15"""

Look at vaccinations for a single date.

 1vac_17_df = vaccines_df[vaccines_df['Date'] == '2021-12-17']
 2print(vac_17_df.iloc[[0,1,2,4,-4,-3,-2,-1], :])
 3
 4"""
 5       Country_Region        Date   Doses_admin  People_partially_vaccinated  People_fully_vaccinated Report_Date_String    UID Province_State 
 6128496    Afghanistan  2021-12-17  5.228706e+06                 4.397449e+06             3.566192e+06         2021-12-18    4.0            NaN 
 7128497        Albania  2021-12-17  2.226267e+06                 1.111494e+06             1.001189e+06         2021-12-18    8.0            NaN 
 8128498        Algeria  2021-12-17  1.229306e+07                 6.875003e+06             5.391232e+06         2021-12-18   12.0            NaN 
 9128500         Angola  2021-12-17  1.080528e+07                 7.246240e+06             3.559041e+06         2021-12-18   24.0            NaN 
10129124          World  2021-12-17  8.647128e+09                 4.402818e+09             3.622553e+09         2021-12-18    NaN            NaN 
11129125          Yemen  2021-12-17  7.860270e+05                 5.566520e+05             3.665870e+05         2021-12-18  887.0            NaN 
12129126         Zambia  2021-12-17  1.342989e+06                 8.066110e+05             9.042010e+05         2021-12-18  894.0            NaN 
13129127       Zimbabwe  2021-12-17  7.058474e+06                 4.023289e+06             3.035185e+06         2021-12-18  716.0            NaN 
14"""

 1fully_vac_17_df = vac_17_df.groupby("Country_Region").sum()[['People_fully_vaccinated']]
 2fully_vac_17_df.rename_axis(None, axis=0, inplace=True)
 3fully_vac_17_df.drop(['World', 'US (Aggregate)'], axis=0, inplace=True)
 4
 5top_10_vac = fully_vac_17_df.sort_values(by = fully_vac_17_df.columns[-1], ascending = False).head(10)
 6top_10_vac
 7
 8"""
 9           People_fully_vaccinated
10China                 1.162488e+09
11India                 5.330402e+08
12US                    2.034792e+08
13Brazil                1.411751e+08
14Indonesia             1.052381e+08
15Japan                 9.822253e+07
16Mexico                6.658651e+07
17Russia                6.207541e+07
18Pakistan              5.936147e+07
19Germany               5.817144e+07
20"""

Countries with highest vaccinations against COVID-19

Load population data

As detailed in Display COVID-19 numbers relative to population the population data is available from the World Bank data. This is available to download in csv, xml or Excel format. The Excel file contains three worksheets with the population data in the first sheet Data.

Access the excel file to determine the sheet names in the file.

1pop_excel_link = "https://api.worldbank.org/v2/en/indicator/SP.POP.TOTL?downloadformat=excel"
2
3pop_xl = pd.ExcelFile(pop_excel_link, engine = "xlrd")
4print(f"Worksheets = {pop_xl.sheet_names}")  # see all sheet names
5
6"""
7Worksheets = ['Data', 'Metadata - Countries', 'Metadata - Indicators']
8"""

The first sheet *Data* contains the population information.

 1# Load the excel worksheet into a dataframe
 2pop_df = pd.read_excel(
 3    pop_excel_link,
 4    engine = "xlrd",
 5    sheet_name = 'Data',
 6    header = 3)
 7
 8print(pop_df.shape)
 9"""
10(266, 65)
11"""
12
13print(pop_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]])
14"""
15                    Country Name Country Code     Indicator Name         2018         2019         2020 
160                          Aruba          ABW  Population, total     105846.0     106310.0     106766.0 
171    Africa Eastern and Southern          AFE  Population, total  643090131.0  660046272.0  677243299.0 
182                    Afghanistan          AFG  Population, total   37171922.0   38041757.0   38928341.0 
19263                 South Africa          ZAF  Population, total   57792520.0   58558267.0   59308690.0 
20264                       Zambia          ZMB  Population, total   17351714.0   17861034.0   18383956.0 
21265                     Zimbabwe          ZWE  Population, total   14438812.0   14645473.0   14862927.0 
22"""

This can be simplified to just read in the last column as we are interested in the latest population.

 1# Load only the desired columns from the excel worksheet into a dataframe
 2pop_2020_df = pd.read_excel(
 3    pop_excel_link,
 4    engine = "xlrd",
 5    sheet_name = 'Data',
 6    usecols = ['Country Name', 'Country Code', '2019'],
 7    header = 3)
 8
 9pop_2020_df.set_index('Country Name', inplace=True)
10pop_2020_df.rename_axis(None, axis=0, inplace=True)
11print(pop_2020_df.shape)
12"""
13(266, 2)
14"""
15
16print(pop_2020_df.iloc[[0,1,2,3,4,5,-5,-4,-3,-2,-1], :])
17"""
18                            Country Code         2019
19Aruba                                ABW     106310.0
20Africa Eastern and Southern          AFE  660046272.0
21Afghanistan                          AFG   38041757.0
22Africa Western and Central           AFW  446911598.0
23Angola                               AGO   31825299.0
24Albania                              ALB    2854191.0
25Kosovo                               XKX    1788878.0
26Yemen, Rep.                          YEM   29161922.0
27South Africa                         ZAF   58558267.0
28Zambia                               ZMB   17861034.0
29Zimbabwe                             ZWE   14645473.0
30"""

There are a number of regions that are aggregate of other countries. These are removed using the data in the second sheet in the excel file (Metadata - Countries).

 1# Need to remove regions that are not countries
 2metadata_df = pd.read_excel(
 3    pop_excel_link,
 4    engine = "xlrd",
 5    sheet_name = 'Metadata - Countries',
 6    header = 0)
 7
 8# metadata_df.isnan('Region')
 9non_country_regions = list(metadata_df[metadata_df['Region'].isna()]['Country Code'])
10print(non_country_regions)
11
12"""
13['AFE', 'AFW', 'ARB', 'CEB', 'CSS', 'EAP', 'EAR', 'EAS', 'ECA', 'ECS', 'EMU', 'EUU', 'FCS', 'HIC', 'HPC', 'IBD', 'IBT', 'IDA', 'IDB', 'IDX', 'LAC', 'LCN', 'LDC', 'LIC', 'LMC', 'LMY', 'LTE', 'MEA', 'MIC', 'MNA', 'NAC', 'OED', 'OSS', 'PRE', 'PSS', 'PST', 'SAS', 'SSA', 'SSF', 'SST', 'TEA', 'TEC', 'TLA', 'TMN', 'TSA', 'TSS', 'UMC', 'WLD']
14"""

These are the top country populations in the world in 2020.

 1# Remove non-countries from the dataframe
 2pop_2020_df = pop_2020_df[~pop_2020_df['Country Code'].isin(non_country_regions)].copy()
 3top_10_pop = pop_2020_df.sort_values(by = pop_2020_df.columns[-1], ascending = False).head(10)
 4
 5print(top_10_pop)
 6"""
 7                   Country Code          2020
 8China                       CHN  1.410929e+09
 9India                       IND  1.380004e+09
10United States               USA  3.294841e+08
11Indonesia                   IDN  2.735236e+08
12Pakistan                    PAK  2.208923e+08
13Brazil                      BRA  2.125594e+08
14Nigeria                     NGA  2.061396e+08
15Bangladesh                  BGD  1.646894e+08
16Russian Federation          RUS  1.441041e+08
17Mexico                      MEX  1.289328e+08
18"""

Top 10 countries by population

Compare countries in the datasets

There are some discrepancies in the countries between the datasets.

 1def extra_countries(df1, df2):
 2    unmatched_covid_countries = sorted(
 3        list(
 4            df1[
 5                ~df1.index
 6                .isin(list(df2.index))
 7            ].index))
 8    return unmatched_covid_countries
 9
10
11def print_list(data, columns=3):
12    # Pretty display for the list of countries
13    for i, x in enumerate(data):
14        print(f"{x.ljust(38, ' ')}", end='')
15        if i%columns==(columns-1):
16            print()

 1extra = extra_countries(deaths_17_df, pop_2020_df)
 2print_list(extra)
 3
 4"""
 5Bahamas                               Brunei                                Burma                                 
 6Congo (Brazzaville)                   Congo (Kinshasa)                      Czechia                               
 7Diamond Princess                      Egypt                                 Gambia                                
 8Holy See                              Iran                                  Korea, South                          
 9Kyrgyzstan                            Laos                                  MS Zaandam                            
10Micronesia                            Russia                                Saint Kitts and Nevis                 
11Saint Lucia                           Saint Vincent and the Grenadines      Slovakia                              
12Summer Olympics 2020                  Syria                                 Taiwan*                               
13US                                    Venezuela                             Yemen    
14"""

 1extra = extra_countries(pop_2020_df, deaths_17_df)
 2print_list(extra)
 3
 4"""
 5American Samoa                        Aruba                                 Bahamas, The                          
 6Bermuda                               British Virgin Islands                Brunei Darussalam                     
 7Cayman Islands                        Channel Islands                       Congo, Dem. Rep.                      
 8Congo, Rep.                           Curacao                               Czech Republic                        
 9Egypt, Arab Rep.                      Faroe Islands                         French Polynesia                      
10Gambia, The                           Gibraltar                             Greenland                             
11Guam                                  Hong Kong SAR, China                  Iran, Islamic Rep.                    
12Isle of Man                           Korea, Dem. People's Rep.             Korea, Rep.                           
13Kyrgyz Republic                       Lao PDR                               Macao SAR, China                      
14Micronesia, Fed. Sts.                 Myanmar                               Nauru                                 
15New Caledonia                         Northern Mariana Islands              Not classified                        
16Puerto Rico                           Russian Federation                    Sint Maarten (Dutch part)             
17Slovak Republic                       St. Kitts and Nevis                   St. Lucia                             
18St. Martin (French part)              St. Vincent and the Grenadines        Syrian Arab Republic                  
19Turkmenistan                          Turks and Caicos Islands              Tuvalu                                
20United States                         Venezuela, RB                         Virgin Islands (U.S.)                 
21Yemen, Rep.                           
22"""

1extra = extra_countries(deaths_17_df, fully_vac_17_df)
2print_list(extra)
3
4"""
5Diamond Princess                      Eritrea                               Holy See                              
6MS Zaandam                            Summer Olympics 2020                  Tonga   
7"""

1extra = extra_countries(fully_vac_17_df, deaths_17_df)
2print_list(extra)
3"""
4empty list
5"""

 1extra = extra_countries(fully_vac_17_df, pop_2020_df)
 2print_list(extra)
 3
 4"""
 5Bahamas                               Brunei                                Burma                                 
 6Congo (Brazzaville)                   Congo (Kinshasa)                      Czechia                               
 7Egypt                                 Gambia                                Iran                                  
 8Korea, South                          Kyrgyzstan                            Laos                                  
 9Micronesia                            Russia                                Saint Kitts and Nevis                 
10Saint Lucia                           Saint Vincent and the Grenadines      Slovakia                              
11Syria                                 Taiwan*                               US                                    
12Venezuela                             Yemen     
13"""

 1extra = extra_countries(pop_2020_df, fully_vac_17_df)
 2print_list(extra)
 3
 4"""
 5American Samoa                        Aruba                                 Bahamas, The                          
 6Bermuda                               British Virgin Islands                Brunei Darussalam                     
 7Cayman Islands                        Channel Islands                       Congo, Dem. Rep.                      
 8Congo, Rep.                           Curacao                               Czech Republic                        
 9Egypt, Arab Rep.                      Eritrea                               Faroe Islands                         
10French Polynesia                      Gambia, The                           Gibraltar                             
11Greenland                             Guam                                  Hong Kong SAR, China                  
12Iran, Islamic Rep.                    Isle of Man                           Korea, Dem. People's Rep.             
13Korea, Rep.                           Kyrgyz Republic                       Lao PDR                               
14Macao SAR, China                      Micronesia, Fed. Sts.                 Myanmar                               
15Nauru                                 New Caledonia                         Northern Mariana Islands              
16Not classified                        Puerto Rico                           Russian Federation                    
17Sint Maarten (Dutch part)             Slovak Republic                       St. Kitts and Nevis                   
18St. Lucia                             St. Martin (French part)              St. Vincent and the Grenadines        
19Syrian Arab Republic                  Tonga                                 Turkmenistan                          
20Turks and Caicos Islands              Tuvalu                                United States                         
21Venezuela, RB                         Virgin Islands (U.S.)                 Yemen, Rep. 
22"""

Create a mapping between the country names in the datasets.

 1unmatched_covid_countries = extra_countries(deaths_17_df, pop_2020_df)
 2
 3# remove cruise ships and disputed teritories that don't have any equivalent country in population dataset
 4unmatched_covid_countries.remove('Diamond Princess')
 5unmatched_covid_countries.remove('MS Zaandam')
 6unmatched_covid_countries.remove('Holy See')
 7unmatched_covid_countries.remove('Taiwan*')
 8unmatched_covid_countries.remove('Summer Olympics 2020')
 9
10# Manually map these countries to the country names used in COVID-19 dataset
11map_pop = ['Bahamas, The', 'Brunei Darussalam', 'Myanmar', 'Congo, Rep.', 'Congo, Dem. Rep.',
12           'Czech Republic', 'Egypt, Arab Rep.', 'Gambia, The', 'Iran, Islamic Rep.', 
13           'Korea, Rep.', 'Kyrgyz Republic', 'Lao PDR', 'Micronesia, Fed. Sts.', 
14           'Russian Federation', 'St. Kitts and Nevis', 'St. Lucia', 'St. Vincent and the Grenadines', 
15           'Slovak Republic', 'Syrian Arab Republic', 'United States', 'Venezuela, RB', 'Yemen, Rep.'
16          ]
17
18country_map = {}
19for c,p in zip(unmatched_covid_countries, map_pop):
20    country_map[p] = c

Mapping of mis-matched country names between Covid-data and Population data:

COVID-19 Country Name	Population Country Name
Bahamas	Bahamas, The
Brunei	Brunei Darussalam
Burma	Myanmar
Congo (Brazzaville)	Congo, Rep.
Congo (Kinshasa)	Congo, Dem. Rep.
Czechia	Czech Republic
Egypt	Egypt, Arab Rep.
Gambia	Gambia, The
Iran	Iran, Islamic Rep.
Korea, South	Korea, Rep.
Kyrgyzstan	Kyrgyz Republic
Laos	Lao PDR
Micronesia	Micronesia, Fed. Sts.
Russia	Russian Federation
Saint Kitts and Nevis	St. Kitts and Nevis
Saint Lucia	St. Lucia
Saint Vincent and the Grenadines	St. Vincent and the Grenadines
Slovakia	Slovak Republic
Syria	Syrian Arab Republic
US	United States
Venezuela	Venezuela, RB
Yemen	Yemen, Rep.

Add a field to the population data to contain the country name.

 1# Add the covid country to the population dataset
 2pop_2020_df['Covid_Country'] = pop_2020_df.apply(
 3    lambda x: country_map[x.name]
 4        if x.name in country_map 
 5        else x.name, 
 6    axis = 1)
 7
 8print(pop_2020_df.iloc[[1,2,3,4,-4,-3,-2,-1], :])
 9
10"""
11             Country Code        2020 Covid_Country
12Afghanistan           AFG  38928341.0   Afghanistan
13Angola                AGO  32866268.0        Angola
14Albania               ALB   2837743.0       Albania
15Andorra               AND     77265.0       Andorra
16Yemen, Rep.           YEM  29825968.0         Yemen
17South Africa          ZAF  59308690.0  South Africa
18Zambia                ZMB  18383956.0        Zambia
19Zimbabwe              ZWE  14862927.0      Zimbabwe
20"""

Confirm that there are no unmatched countries in the Covid datasets.

 1print(list(deaths_17_df[
 2    ~deaths_17_df.index
 3    .isin(list(pop_2020_df['Covid_Country']))
 4].index))
 5
 6"""
 7['Diamond Princess', 'Holy See', 'MS Zaandam', 'Summer Olympics 2020', 'Taiwan*']
 8"""
 9
10print(list(fully_vac_17_df[
11    ~fully_vac_17_df.index
12    .isin(list(pop_2020_df['Covid_Country']))
13].index))
14
15"""
16['Taiwan*']
17"""

Merge datasets on country

A new dataframe is created with the data for the deaths from covid-19 and the vaccinations as well as the population for each country.

 1covid_pop_df = deaths_17_df.merge(
 2    fully_vac_17_df,
 3    how='left',
 4    left_index = True,
 5    right_on = fully_vac_17_df.index
 6).merge (
 7    pop_2020_df,
 8    how='left',
 9    left_on = covid_df.index,
10    right_on = 'Covid_Country'
11)
12
13covid_pop_df.set_index('key_0', inplace=True)
14covid_pop_df.rename_axis(None, axis=0, inplace=True)
15
16# # Rename column names for deaths and population
17covid_pop_df = covid_pop_df.rename(columns={'12/17/21': 'Deaths'})
18covid_pop_df = covid_pop_df.rename(columns={'2020': 'Population'})
19
20# # Remove unneeded columns
21covid_pop_df.drop(['Country Code', 'Covid_Country'], axis=1, inplace=True)
22
23print(covid_pop_df.iloc[[1,2,3,4,-4,-3,-2,-1], :])
24
25"""
26                    Deaths  People_fully_vaccinated  Population
27Albania               3158                1001189.0   2837743.0
28Algeria               6175                5391232.0  43851043.0
29Andorra                134                  50551.0     77265.0
30Angola                1738                3559041.0  32866268.0
31West Bank and Gaza    4855                      0.0   4803269.0
32Yemen                 1974                 366587.0  29825968.0
33Zambia                3674                 904201.0  18383956.0
34Zimbabwe              4779                3035185.0  14862927.0
35"""

Overview of Countries with highest deaths from covid per capita

The number of people vaccinated is calculated per 1,000 of the population and the number of people died is calculated per 100,000 of the population. The top countries with the highest deaths per capita and those with the highest vaccinations are shown for countries with populations greater than 1 million.

 1covid_pop_df['deaths_100k'] = covid_pop_df.apply(lambda x: (x.Deaths / x.Population) * 100000, axis = 1)
 2
 3covid_pop_df['vac_100'] = covid_pop_df.apply(lambda x: (x.People_fully_vaccinated / x.Population) * 100, axis = 1)
 4
 5print(covid_pop_df.iloc[[1,2,3,4,-4,-3,-2,-1], :])
 6
 7"""
 8                    Deaths  People_fully_vaccinated  Population  deaths_100k    vac_100  
 9Albania               3158                1001189.0   2837743.0   111.285624  35.281172  
10Algeria               6175                5391232.0  43851043.0    14.081763  12.294421  
11Andorra                134                  50551.0     77265.0   173.429108  65.425484  
12Angola                1738                3559041.0  32866268.0     5.288097  10.828857  
13West Bank and Gaza    4855                      0.0   4803269.0   101.076996   0.000000  
14Yemen                 1974                 366587.0  29825968.0     6.618394   1.229087  
15Zambia                3674                 904201.0  18383956.0    19.984817   4.918425  
16Zimbabwe              4779                3035185.0  14862927.0    32.153828  20.421179  
17"""

 1top_10 = (covid_pop_df[covid_pop_df['Population'] > 1000000]
 2          .sort_values(by = ['deaths_100k'], ascending = False)
 3          .head(20))
 4
 5bg_color = 'rgba(208,225,242,1.0)'
 6
 7fig = go.Figure(
 8    data = [go.Bar(x = list(top_10["deaths_100k"]),
 9                   y = list(top_10.index),
10                   hovertemplate = "%{y}: <br>Deaths per 100K: %{x:.0f}<extra></extra>",
11                   orientation = "h")]
12)
13
14fig.update_layout(
15    # Set default font
16        font = dict(
17        family = "Droid Sans",
18        size = 14
19    ),
20    # Set figure title
21    title = dict(
22        text = f"<b>Countries with highest deaths from Covid-19 per 100K of population [Dec 17, 2020]</b>",
23        xref = 'container',
24        yref = 'container',
25        x = 0.5,
26        y = 0.92,
27        xanchor = 'center',
28        yanchor = 'middle',
29        font = dict(family = 'Droid Sans', size = 24)
30    ),
31    # set the plot bacground color
32    plot_bgcolor = bg_color,
33    # set the hover background color
34    hoverlabel = dict(
35        bgcolor = 'rgba(75,152,201,0.2)',
36        font_size = 16
37    ),
38    paper_bgcolor = bg_color,
39)
40
41fig.update_traces(marker_color = 'rgb(75,152,201)')
42fig['layout']['yaxis']['autorange'] = "reversed"
43fig.show()

Countries with highest deaths from Covid-19 per 100K of population

These are the countries with populations over 1 million with the lowest number of deaths from covid-19 per capita.

Countries with lowest deaths from Covid-19 per 100K of population

These are the countries with populations over 1 million with the highest percent of the population vaccinated against covid-19.

Countries with lowest deaths from Covid-19 per 100K of population

Vaccination status in countries with highest deaths

A bar plot in relative mode is used to show percent vaccinated with number of deaths in countries with the highest deaths from cvid-19.

 1top_20 = (covid_pop_df[covid_pop_df['Population'] > 1000000]
 2          .sort_values(by = ['deaths_100k'], ascending = False)
 3          .head(20))
 4
 5neg_vac = [(-1 * x) for x in top_20.vac_100]
 6
 7bg_color = 'rgba(208,225,242,1.0)'
 8
 9fig = go.Figure()
10fig.add_trace(go.Bar(x = [v*6 for v in neg_vac],
11                     y = list(top_20.index),
12                     marker_color = 'rgb(30,75,80)',
13                     name = 'vaccinations',
14                     customdata = [(v * -1) for v in neg_vac],
15                     hovertemplate = "%{y}: <br>Vaccinated: %{customdata:.0f}%<extra></extra>",
16                     orientation = "h"))
17fig.add_trace(go.Bar(x = list(top_20["deaths_100k"]),
18                     y = list(top_20.index),
19                     marker_color = 'rgb(75,152,201)',
20                     name = 'deaths',
21                     hovertemplate = "%{y}: <br>Deaths per 100K: %{x:.0f}<extra></extra>",
22                     orientation = "h"))
23
24
25fig.update_layout(
26    # Set default font
27    font = dict(
28        family = "Droid Sans",
29        size = 14
30    ),
31    # Set figure title
32    title = dict(
33        text = f"<b>Countries with highest deaths from Covid-19 per 100K of population [Dec 17, 2020]</b>",
34        xref = 'container',
35        yref = 'container',
36        x = 0.5,
37        y = 0.95,
38        xanchor = 'center',
39        yanchor = 'middle',
40        font = dict(family = 'Droid Sans', size = 24)
41    ),
42    # set x-axis
43    xaxis = dict(
44        title = '',
45        showticklabels = False,
46    ),
47    
48    legend=dict(
49        orientation="h",
50        yanchor="bottom",
51        y=1.02,
52        xanchor="center",
53        x=0.41
54    ),
55
56    # set the plot bacground color
57    plot_bgcolor = bg_color,
58    # set the hover background color
59    hoverlabel = dict(
60        bgcolor = 'rgba(75,152,201,0.2)',
61        font_size = 16
62    ),
63    paper_bgcolor = bg_color,
64)
65
66fig.update_layout(barmode='relative')
67fig['layout']['yaxis']['autorange'] = "reversed"
68fig.show()

Countries with highest deaths from Covid-19 per 100K of population with percentage vaccinated

Conclusion

The covid-19 pandemic is far from over and their continues to be a number of Country status for Covid and Vaccinations. There is a trend that those countries with higher percentage of the population vaccinated have lower deaths per capita, but there are exceptions to this.

Daily counts and rolling average of COVID-19 numbers

Eric Callanan — Sunday, 20 Dec 2020

The daily counts of confirmed cases or deaths from COVID-19 can fluctuate greatly from day to day. This can be down to time factors such as week day vs weekend or delays in testing or gathering and collating results. A rolling average is a calculated average of different subsets of the full data set. It is used to smooth out fluctuations in the data.

Data sources:
The data used in this article is retrieved from the Johns Hopkins University who have made the data available on GitHub. More information about COVID-19 and the coronavirus is available from Coronavirus disease (COVID-19) advice for the public.

Load COVID-19 data into dataframes

There are two sets of data to load the global confirmed cases and the global deaths.

time_series_covid19_deaths_global.csv
time_series_covid19_confirmed_global.csv

The data can be loaded directly from the GitHub page as in the code below. The data is cleaned to remove unwanted fields and group the data by country with the following function.

 1def load_clean_covid_data(csv_path):
 2    df = pd.read_csv(csv_path)
 3
 4    # 1. Drop unwanted columns
 5    df.drop(['Province/State', 'Lat', 'Long'], axis = 1, inplace = True)
 6
 7    # 2. Group by Country
 8    df = df.groupby('Country/Region').sum()
 9
10    # 3. Convert column headings to datetime
11    df.columns = pd.to_datetime(df.columns)
12
13    # 4. Remove the column name
14    df.rename_axis(None, axis = 0, inplace = True)
15
16    return df

Load the data for Confirmed cases and the data for Deaths from COVID-19 into dataframes.

 1
 2confirmed_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv')
 3confirmed_df.shape
 4"""
 5(191, 333)
 6"""
 7
 8confirmed_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
 9"""
10             2020-01-22  2020-01-23  2020-01-24  2020-12-17  2020-12-18  2020-12-19  
11Afghanistan           0           0           0       49378       49621       49681  
12Albania               0           0           0       51424       52004       52542  
13Algeria               0           0           0       93933       94371       94781  
14Yemen                 0           0           0        2087        2087        2087  
15Zambia                0           0           0       18504       18575       18620  
16Zimbabwe              0           0           0       11866       12047       12151  
17"""
18
19
20deaths_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv')
21deaths_df.shape
22"""
23(191, 333)
24"""
25
26deaths_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
27"""
28             2020-01-22  2020-01-23  2020-01-24  2020-12-17  2020-12-18  2020-12-19  
29Afghanistan           0           0           0        2025        2030        2047  
30Albania               0           0           0        1055        1066        1074  
31Algeria               0           0           0        2640        2647        2659  
32Yemen                 0           0           0         606         606         606  
33Zambia                0           0           0         369         373         373  
34Zimbabwe              0           0           0         314         316         318  
35"""

Get the top 10 countries

Get the names of the top ten countries based on the highest deaths for the latest date.

 1top_countries = list(deaths_df
 2                 .iloc[:, -1]
 3                 .sort_values(ascending = False)
 4                 .head(10)
 5                 .index)
 6
 7top_countries
 8"""
 9['US',
10 'Brazil',
11 'India',
12 'Mexico',
13 'Italy',
14 'United Kingdom',
15 'France',
16 'Iran',
17 'Russia',
18 'Spain']
19"""

Display the deaths for a country over time

The data in time_series_covid19_deaths_global.csv is the cumulative number of deaths from COVID-19 recorded in each country every day. This is displayed in a line chart.

 1c = 'US'
 2data = deaths_df.loc[c]
 3date = deaths_df.columns[-1].strftime('%b %d %Y')
 4
 5bg_color = 'rgba(208, 225, 242, 1.0)'
 6line_color = 'rgba(75, 152, 201, 1.0)'
 7grid_color = 'rgba(75, 152, 201, 0.3)'
 8
 9fig = go.Figure()
10fig.add_trace(go.Scatter(
11    x = data.index,
12    y = data,
13    name = 'US',
14    hovertemplate = [f"""<b>{c}</b><br>
15<br>Deaths: {x:,.0F}
16<extra></extra>
17""" for x in data] ))
18
19fig.update_layout(
20    # Set figure title
21    title = dict(
22        text = f'Deaths from COVID-19 in <b>"{c}"</b> <br> {date}',
23        xref = 'container',
24        yref = 'container',
25        x = 0.5,
26        y = 0.9,
27        xanchor = 'center',
28        yanchor = 'middle',
29        font = dict(family = 'Droid Sans', size = 28)
30    ),
31    # set x-axis
32    xaxis = dict(
33        title = 'Date',
34        linecolor = line_color,
35        linewidth = 2,
36        gridcolor = grid_color,
37        showticklabels = True,
38        ticks = 'outside',
39    ),
40    # set y-axis
41    yaxis = dict(
42        title = 'Number of deaths',
43        rangemode = 'tozero',
44        linecolor = line_color,
45        linewidth = 2,
46        gridcolor = grid_color,
47        showticklabels = True,
48        ticks = 'outside',
49    ),
50    # set the plot bacground color
51    plot_bgcolor = bg_color,
52    paper_bgcolor = bg_color,
53)
54
55fig.show()

Total deaths from COVID-19 in US - December 19 2020

Display daily counts for a single country

Reduce the dataset by selecting the data for the top 10 countries. The daily count is easily got using pandas.DataFrame.diff function. This calculates the diference of data elements from another element in the dataframe. The default is to calculate the difference between the element in the previous row, but an axis parameter of 1 or columns can be used to calculate the difference between the element in the previous column.

 1top_df = deaths_df.loc[top_countries]
 2top_daily_df = top_df.diff(axis=1)
 3
 4top_daily_df.iloc[:, -10:]
 5"""
 6                2020-12-10  2020-12-11  2020-12-12  2020-12-13  2020-12-14  2020-12-15  2020-12-16  2020-12-17  2020-12-18  2020-12-19  
 7US                  2920.0      3283.0      2354.0      1389.0      1484.0      2984.0      3668.0      3345.0      2814.0      2571.0  
 8Brazil               770.0       672.0       686.0       279.0       433.0       964.0       936.0      1092.0       823.0       706.0  
 9India                413.0       443.0       391.0       336.0       354.0       387.0       355.0       338.0       347.0       341.0  
10Mexico               671.0       693.0       685.0       249.0       345.0       801.0       670.0       718.0       762.0       627.0  
11Italy                887.0       761.0       649.0       484.0       491.0       846.0       680.0       683.0       674.0       553.0  
12United Kingdom       516.0       424.0       520.0       144.0       233.0       506.0       612.0       532.0       490.0       537.0  
13France               296.0       628.0        -1.0       344.0       376.0       791.0       290.0       261.0       612.0       189.0  
14Iran                 284.0       231.0       222.0       247.0       251.0       223.0       213.0       212.0       178.0       175.0  
15Russia               549.0       601.0       553.0       481.0       442.0       564.0       584.0       574.0       602.0       574.0  
16Spain                325.0       280.0         0.0         0.0       389.0       388.0       195.0       181.0       149.0         0.0  
17"""

Table: Deaths from COVID-19 in previous 7 days in countries with highest deaths on December 19 2020

Country	Dec 13	Dec 14	Dec 15	Dec 16	Dec 17	Dec 18	Dec 19
US	1,389	1,484	2,984	3,668	3,345	2,814	2,571
Brazil	279	433	964	936	1,092	823	706
India	336	354	387	355	338	347	341
Mexico	249	345	801	670	718	762	627
Italy	484	491	846	680	683	674	553
United Kingdom	144	233	506	612	532	490	537
France	344	376	791	290	261	612	189
Iran	247	251	223	213	212	178	175
Russia	481	442	564	584	574	602	574
Spain	0	389	388	195	181	149	0

The daily counts is plotted for a single country.

 1c = 'US'
 2data = top_daily_df.loc[c]
 3date = top_daily_df.columns[-1].strftime('%b %d %Y')
 4
 5bg_color = 'rgba(208, 225, 242, 1.0)'
 6line_color = 'rgba(75, 152, 201, 1.0)'
 7grid_color = 'rgba(75, 152, 201, 0.3)'
 8
 9fig = go.Figure()
10fig.add_trace(go.Scatter(
11    x = data.index,
12    y = data,
13    name = 'US',
14    hovertemplate = [f"""<b>{c}</b><br>
15<br>Deaths: {x:,.0F}
16<extra></extra>
17""" for x in data] ))
18
19fig.update_layout(
20    # Set figure title
21    title = dict(
22        text = f'Daily deaths from COVID-19 in <b>"{c}"</b> <br> {date}',
23        xref = 'container',
24        yref = 'container',
25        x = 0.5,
26        y = 0.9,
27        xanchor = 'center',
28        yanchor = 'middle',
29        font = dict(family = 'Droid Sans', size = 28)
30    ),
31    # set x-axis
32    xaxis = dict(
33        title = 'Date',
34        linecolor = line_color,
35        linewidth = 2,
36        gridcolor = grid_color,
37        showticklabels = True,
38        ticks = 'outside',
39    ),
40    # set y-axis
41    yaxis = dict(
42        title = 'Number of deaths',
43        rangemode = 'tozero',
44        linecolor = line_color,
45        linewidth = 2,
46        gridcolor = grid_color,
47        showticklabels = True,
48        ticks = 'outside',
49    ),
50    # set the plot bacground color
51    plot_bgcolor = bg_color,
52    paper_bgcolor = bg_color,
53)
54
55fig.show()

Daily deaths from COVID-19 in US - December 19 2020

Display rolling average on daily counts

The daily counts fluctuate greatly from day to day and it can be difficult to see the overall trend. A rolling average calculates an average of different subsets of the full data set. It is used to smooth out fluctuations in the data. In this case the pandas.Series.rolling function is used to create a number of rolling average to determine the best value to use.

Here is an example of calculating the rolling average of 3 days for the latest 20 days on the number of deaths from COVID-19 in the US. The chart is used to determine the best number of days to set for the rolling window. Rolling window sizes of 3, 5, 7 and 10 were compared. The different results can be switched on and off in the Plotly chart and areas of the chart can be zoomed in on. These results show that 7 is the best number to select for the rolling window size. This is likely due to daily fluctuations based on weekday that are smoothed out when taking the week on average.

 1c = 'US'
 2data = top_daily_df.loc[c].tail(20)
 3rolling = data.rolling(3).mean()
 4
 5data
 6"""
 7['1153.0', '2588.0', '2775.0', '2843.0', '2594.0', '2257.0', '1122.0',
 8'1419.0', '2520.0', '3064.0', '2920.0', '3283.0', '2354.0', '1389.0',
 9'1484.0', '2984.0', '3668.0', '3345.0', '2814.0', '2571.0']
10"""
11
12rolling
13"""
14['NAN', 'NAN', '2172.0', '2735.3', '2737.3', '2564.7', '1991.0',
15'1599.3', '1687.0', '2334.3', '2834.7', '3089.0', '2852.3', '2342.0',
16'1742.3', '1952.3', '2712.0', '3332.3', '3275.7', '2910.0']
17"""

 1c = 'US'
 2data = top_daily_df.loc[c]
 3date = top_daily_df.columns[-1].strftime('%b %d %Y')
 4r3 = data.rolling(3).mean()
 5r5 = data.rolling(5).mean()
 6r7 = data.rolling(7).mean()
 7r10 = data.rolling(10).mean()
 8
 9bg_color = 'rgba(208, 225, 242, 1.0)'
10line_color = 'rgba(75, 152, 201, 1.0)'
11grid_color = 'rgba(75, 152, 201, 0.3)'
12
13fig = go.Figure()
14for d, n in zip([data, r3, r5, r7, r10], [1, 3, 5, 7, 10]):    
15    fig.add_trace(go.Scatter(
16        x = d.index,
17        y = d,
18        name = f'rolling window = "{n}"',
19        hovertemplate = [f"""<b>{c}</b>
20    <br>Rolling window = {n}<br>
21    <br>Deaths: {x:,.0F}
22    <extra></extra>
23    """ for x in d] ))
24
25fig.update_layout(
26    # Set figure title
27    title = dict(
28        text = f'Daily deaths from COVID-19 in <b>"{c}"</b> <br> {date}',
29        xref = 'container',
30        yref = 'container',
31        x = 0.5,
32        y = 0.9,
33        xanchor = 'center',
34        yanchor = 'middle',
35        font = dict(family = 'Droid Sans', size = 28)
36    ),
37    # set x-axis
38    xaxis = dict(
39        title = 'Date',
40        linecolor = line_color,
41        linewidth = 2,
42        gridcolor = grid_color,
43        showticklabels = True,
44        ticks = 'outside',
45    ),
46    # set y-axis
47    yaxis = dict(
48        title = 'Number of deaths',
49        rangemode = 'tozero',
50        linecolor = line_color,
51        linewidth = 2,
52        gridcolor = grid_color,
53        showticklabels = True,
54        ticks = 'outside',
55    ),
56    # set the plot bacground color
57    plot_bgcolor = bg_color,
58    paper_bgcolor = bg_color,
59)
60
61fig.show()

Daily deaths from COVID-19 in US showing means for different rolling window sizes - December 19 2020

Identify the best rolling average window size

It can be easier to identify the better rolling window size by plotting the different options on separate plots. This confirms that the best rolling window setting is for 7 days. The 7-day rolling average smooths out the daily fluctuations and makes it easier to see trends in the data. This confirmed by creating a number of charts for different countries.

 1c = 'US'
 2data = top_daily_df.loc[c]
 3date = top_daily_df.columns[-1].strftime('%b %d %Y')
 4r3 = data.rolling(3).mean()
 5r5 = data.rolling(5).mean()
 6r7 = data.rolling(7).mean()
 7r10 = data.rolling(10).mean()
 8r20 = data.rolling(10).mean()
 9
10bg_color = 'rgba(208, 225, 242, 1.0)'
11line_color = 'rgba(75, 152, 201, 1.0)'
12grid_color = 'rgba(75, 152, 201, 0.3)'
13
14fig = make_subplots(
15    rows = 3,
16    cols = 2,
17    subplot_titles = (
18        "Daily Data",
19        "Rolling Window 3",
20        "Rolling Window 5",
21        "Rolling Window 7",
22        "Rolling Window 10",
23        "Rolling Window 20"))
24
25fig.add_trace(go.Scatter(x = data.index, y = data), row = 1, col = 1)
26fig.add_trace(go.Scatter(x = r3.index, y = r3), row = 1, col = 2)
27fig.add_trace(go.Scatter(x = r5.index, y = r5), row = 2, col = 1)
28fig.add_trace(go.Scatter(x = r7.index, y = r7), row = 2, col = 2)
29fig.add_trace(go.Scatter(x = r10.index, y = r10), row = 3, col = 1)
30fig.add_trace(go.Scatter(x = r20.index, y = r20), row = 3, col = 2)
31
32fig.update_layout(
33    showlegend=False
34)
35
36fig.show()

Deaths from COVID-19 in US showing different rolling averages in multiple subplots

Deaths from COVID-19 in UK showing different rolling averages in multiple subplots

Display rolling averages for top ten countries

Display the 7-day rolling average of Deaths from COVID-19 for the top ten countries with the highest deaths on December 19 2020.

 1date = top_daily_df.columns[-1].strftime('%b %d %Y')
 2
 3bg_color = 'rgba(208, 225, 242, 1.0)'
 4line_color = 'rgba(75, 152, 201, 1.0)'
 5grid_color = 'rgba(75, 152, 201, 0.3)'
 6
 7fig = go.Figure()
 8for i, row in top_daily_df.iterrows():
 9    r7 = row.rolling(7).mean()
10    fig.add_trace(go.Scatter(
11        x = r7.index,
12        y = r7,
13        name = i,
14        hovertemplate = [f"""<b>{r7.name}</b>
15    <br>Deaths: {x:,.0F}
16    <extra></extra>
17    """ for x in r7]))
18
19fig.update_layout(
20    # Set figure title
21    title = dict(
22        text = f'7-day daily average of deaths from COVID-19 in top ten countries<br> {date}',
23        xref = 'container',
24        yref = 'container',
25        x = 0.5,
26        y = 0.9,
27        xanchor = 'center',
28        yanchor = 'middle',
29        font = dict(family = 'Droid Sans', size = 24)
30    ),
31    # set x-axis
32    xaxis = dict(
33        title = 'Date',
34        linecolor = line_color,
35        linewidth = 2,
36        gridcolor = grid_color,
37        showticklabels = True,
38        ticks = 'outside',
39    ),
40    # set y-axis
41    yaxis = dict(
42        title = '7-day daily average of deaths',
43        rangemode = 'tozero',
44        linecolor = line_color,
45        linewidth = 2,
46        gridcolor = grid_color,
47        showticklabels = True,
48        ticks = 'outside',
49    ),
50    # set the plot bacground color
51    plot_bgcolor = bg_color,
52    paper_bgcolor = bg_color,
53)
54
55fig.show()

7-day average of Deaths per day from COVID-19 in countries with highest deaths

A similar plot can be created for confirmed cases for the same ten countries with highest deaths.

7-day average of Confirmed Cases per day of COVID-19 in countries with highest deaths

Display rolling averages and the daily counts

Display the 7-day rolling average of Deaths from COVID-19 as well as the actual daily counts in the same chart. The daily counts are shown as bars and the rolling average is a line. The results for the countries with the highest deaths is displayed. The plots are separated per country. There is a lot of data collected to date and sections of the charts can be chosen to zoom in on a particular time period. Another approach is to just plot the data for the most recent 30 days to give a picture of what the current trend is.

 1date = top_daily_df.columns[-1].strftime('%b %d %Y')
 2
 3bg_color = 'rgba(208, 225, 242, 1.0)'
 4line_color = 'rgba(75, 152, 201, 1.0)'
 5grid_color = 'rgba(75, 152, 201, 0.3)'
 6
 7fig = make_subplots(
 8    rows = 5,
 9    cols = 2,
10    shared_xaxes = True,
11    subplot_titles = (top_daily_df.index)
12    )
13
14j = 0
15for i, row in top_daily_df.iterrows():
16    r7 = row.rolling(7).mean()
17    fig.add_trace(go.Bar(x = row.index, y = row, name = i), row = j%5+1, col = j%2+1)
18    fig.add_trace(go.Scatter(x = r7.index, y = r7, name = i), row = j%5+1, col = j%2+1)
19
20    j += 1
21
22
23fig.update_xaxes(dict(
24        linecolor = line_color,
25        gridcolor = grid_color,
26    ))
27
28fig.update_yaxes(dict(
29        rangemode = 'tozero',
30        linecolor = line_color,
31        gridcolor = grid_color,
32    ))
33
34fig.update_layout(
35    # Set figure title
36    title = dict(
37        text = f'Daily deaths from COVID-19 in <b>"{c}"</b> <br> {date}',
38        xref = 'container',
39        yref = 'container',
40        x = 0.5,
41        y = 0.95,
42        xanchor = 'center',
43        yanchor = 'middle',
44        font = dict(family = 'Droid Sans', size = 22)
45    ),
46    showlegend=False,
47    # set the plot bacground color
48    plot_bgcolor = bg_color,
49    paper_bgcolor = bg_color,
50)
51
52fig.show()

Daily counts and 7-day average of Deaths from COVID-19 in countries with highest deaths

Daily counts and 7-day average of Deaths from COVID-19 for latest 30 days

Similar charts can be created for the daily counts and rolling average of confirmed cases.

Daily counts and 7-day average of Confirmed Cases of COVID-19 in countries with highest deaths

Daily counts and 7-day average of Confirmed Cases of COVID-19 for latest 30 days

Conclusion

Once again calculations can be applied to a whole dataset easily with Pandas. The diff function is used to easily convert the running totals to daily differences and the rolling function is used to group a subset of data together and calculate a running average. The running average helps smooth the day to day fluctuations in the data and makes it easier to spot a trend.

Plotly is an open-source graphing library for Python that produces interactive charts.

Display COVID-19 numbers relative to population

Eric Callanan — Sunday, 13 Dec 2020

The total number of cases of COVID-19 and deaths from COVID-19 in a country will vary based on a number of factors. It does not seem right to compare absolute numbers in different countries, where the populations of those countries varies greatly. This article will show how to take the absolute numbers and express the confirmed cases and deaths per 100,000 of the population and thus allow comparison between countries.

Load COVID-19 data into dataframes

There are two sets of data to load the global confirmed cases and the global deaths.

time_series_covid19_deaths_global.csv
time_series_covid19_confirmed_global.csv

The data can be loaded directly from the GitHub page as in the code below. The data is cleaned to remove unwanted fields and group the data by country with the following function.

 1def load_clean_covid_data(csv_path):
 2    df = pd.read_csv(csv_path)
 3
 4    # 1. Drop unwanted columns
 5    df.drop(['Province/State', 'Lat', 'Long'], axis = 1, inplace = True)
 6
 7    # 2. Group by Country
 8    df = df.groupby('Country/Region').sum()
 9
10    # 3. Convert column headings to datetime
11    df.columns = pd.to_datetime(df.columns)
12
13    # 4. Remove the column name
14    df.rename_axis(None, axis = 0, inplace = True)
15
16    return df

Load the data for Confirmed cases and the data for Deaths from COVID-19 into dataframes.

 1
 2confirmed_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv')
 3confirmed_df.shape
 4"""
 5(191, 326)
 6"""
 7
 8confirmed_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
 9"""
10             2020-01-22  2020-01-23  2020-01-24  2020-12-10  2020-12-11  2020-12-12
11Afghanistan           0           0           0       48053       48116       48229
12Albania               0           0           0       46061       46863       47742
13Algeria               0           0           0       90579       91121       91638
14Yemen                 0           0           0        2081        2082        2083
15Zambia                0           0           0       18091       18161       18217
16Zimbabwe              0           0           0       11081       11162       11219
17"""
18
19
20deaths_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv')
21deaths_df.shape
22"""
23(191, 326)
24"""
25
26deaths_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
27"""
28             2020-01-22  2020-01-23  2020-01-24  2020-12-10  2020-12-11  2020-12-12  
29Afghanistan           0           0           0        1935        1945        1956  
30Albania               0           0           0         965         977         989  
31Algeria               0           0           0        2564        2575        2584  
32Yemen                 0           0           0         606         606         606  
33Zambia                0           0           0         364         365         366  
34Zimbabwe              0           0           0         305         306         307  
35"""

Load global population data into dataframe

World population data is available from the World Bank data. This is available to download in csv, xml or Excel format. The Excel file contains three worksheets with the population data in the first sheet Data. This is loaded into a dataframe.

1pop_xl = pd.ExcelFile("/tmp/API_SP.POP.TOTL_DS2_en_excel_v2_1740169.xls", engine = "xlrd")
2pop_xl.sheet_names
3"""
4['Data', 'Metadata - Countries', 'Metadata - Indicators']
5"""

 1# Load the excel worksheet into a dataframe
 2pop_df = pd.read_excel(
 3    "/tmp/API_SP.POP.TOTL_DS2_en_excel_v2_1740169.xls",
 4    engine = "xlrd",
 5    sheet_name = 'Data',
 6    header = 3)
 7
 8pop_df.shape
 9"""
10(264, 65)
11"""
12
13pop_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
14"""
15     Country Name Country Code     Indicator Name        2018        2019  2020  
160           Aruba          ABW  Population, total    105845.0    106314.0   NaN  
171     Afghanistan          AFG  Population, total  37172386.0  38041754.0   NaN  
182          Angola          AGO  Population, total  30809762.0  31825295.0   NaN  
19261  South Africa          ZAF  Population, total  57779622.0  58558270.0   NaN  
20262        Zambia          ZMB  Population, total  17351822.0  17861030.0   NaN  
21263      Zimbabwe          ZWE  Population, total  14439018.0  14645468.0   NaN  
22"""

Extract population data for 2019 into a dataframe.

 1# Get population for countries for 2019
 2pop_2019_df = pop_df[['Country Name', '2019']].copy()
 3
 4pop_2019_df.shape
 5"""
 6(264, 2)
 7"""
 8
 9pop_2019_df.iloc[[0,1,2,3,4,-5,-4-3,-2,-1], :]
10"""
11    Country Name          2019
120          Aruba  1.063140e+05
131    Afghanistan  3.804175e+07
142         Angola  3.182530e+07
153        Albania  2.854191e+06
164        Andorra  7.714200e+04
17259       Kosovo  1.794248e+06
18257        World  7.673534e+09
19262       Zambia  1.786103e+07
20263     Zimbabwe  1.464547e+07
21"""

Load population data for a single year

After exploring the population dataset, the code above can be simplified to just load data for a single year. The population data for the year 2019 can be loaded directly from the Excel file. The Country Code cannot be used, as it is not present in the COVID-19 data.

 1# Load country and data for 2019 from the excel worksheet into a dataframe
 2pop_2019_df = pd.read_excel(
 3    "/tmp/API_SP.POP.TOTL_DS2_en_excel_v2_1740169.xls",
 4    engine = "xlrd",
 5    sheet_name = 'Data',
 6    usecols = ['Country Name', '2019'],
 7    header = 3)
 8
 9pop_2019_df.shape
10"""
11(264, 2)
12"""
13
14pop_2019_df.iloc[[0,1,2,3,4,-5,-4-3,-2,-1], :]
15"""
16    Country Name          2019
170          Aruba  1.063140e+05
181    Afghanistan  3.804175e+07
192         Angola  3.182530e+07
203        Albania  2.854191e+06
214        Andorra  7.714200e+04
22259       Kosovo  1.794248e+06
23257        World  7.673534e+09
24262       Zambia  1.786103e+07
25263     Zimbabwe  1.464547e+07
26"""

Compare countries in COVID-19 datasets with population dataset

Examine the matching of country names between population data and COVID-19 data as country codes are not used. This shows that there are 25 "Country/Region" that are not present in the population data.

 1# Countries in confirmed_df dataframe that are not in pop_2019_df
 2unmatched_covid_countries = sorted(
 3    list(
 4        confirmed_df[
 5            ~confirmed_df.index
 6            .isin(list(pop_2019_df['Country Name']))
 7        ].index))
 8
 9len(unmatched_covid_countries)
10"""
1125
12"""
13
14unmatched_covid_countries
15"""
16Bahamas                               Brunei                                Burma                                 
17Congo (Brazzaville)                   Congo (Kinshasa)                      Czechia                               
18Diamond Princess                      Egypt                                 Gambia                                
19Holy See                              Iran                                  Korea, South                          
20Kyrgyzstan                            Laos                                  MS Zaandam                            
21Russia                                Saint Kitts and Nevis                 Saint Lucia                           
22Saint Vincent and the Grenadines      Slovakia                              Syria                                 
23Taiwan*                               US                                    Venezuela                             
24Yemen                                 
25"""

Four of these "Country/Region" cannot be mapped to a country in the population data. Two of these are cruise ships, Taiwan* is a disputed territory and the Holy See may be included in smal states. These are removed from the list of unmatched countries.

 1unmatched_covid_countries.remove('Diamond Princess')
 2unmatched_covid_countries.remove('MS Zaandam')
 3unmatched_covid_countries.remove('Holy See')
 4unmatched_covid_countries.remove('Taiwan*')
 5
 6len(unmatched_covid_countries)
 7"""
 821
 9"""
10
11unmatched_covid_countries
12"""
13Bahamas                               Brunei                                Burma                                 
14Congo (Brazzaville)                   Congo (Kinshasa)                      Czechia                               
15Egypt                                 Gambia                                Iran                                  
16Korea, South                          Kyrgyzstan                            Laos                                  
17Russia                                Saint Kitts and Nevis                 Saint Lucia                           
18Saint Vincent and the Grenadines      Slovakia                              Syria                                 
19US                                    Venezuela                             Yemen                                 
20"""

The remaining "Country/Region" are manually mapped to Countries in the population dataset.

COVID-19 Country Name	Population Country Name
Bahamas	Bahamas, The
Brunei	Brunei Darussalam
Burma	Myanmar
Congo (Brazzaville)	Congo, Rep.
Congo (Kinshasa)	Congo, Dem. Rep.
Czechia	Czech Republic
Egypt	Egypt, Arab Rep.
Gambia	Gambia, The
Iran	Iran, Islamic Rep.
Korea, South	Korea, Rep.
Kyrgyzstan	Kyrgyz Republic
Laos	Lao PDR
Russia	Russian Federation
Saint Kitts and Nevis	St. Kitts and Nevis
Saint Lucia	St. Lucia
Saint Vincent and the Grenadines	St. Vincent and the Grenadines
Slovakia	Slovak Republic
Syria	Syrian Arab Republic
US	United States
Venezuela	Venezuela, RB
Yemen	Yemen, Rep.

Update country names in population dataset

A new column is added to the population dataset to contain the country name from the COVID-19 dataset if it is different or the original country if there is no difference. This could update the country name in place, but it is better to have the two names for verification.

 1map_pop = ['Bahamas, The', 'Brunei Darussalam', 'Myanmar', 'Congo, Rep.', 'Congo, Dem. Rep.',
 2           'Czech Republic', 'Egypt, Arab Rep.', 'Gambia, The', 'Iran, Islamic Rep.',
 3           'Korea, Rep.', 'Kyrgyz Republic', 'Lao PDR', 'Russian Federation', 'St. Kitts and Nevis',
 4           'St. Lucia', 'St. Vincent and the Grenadines', 'Slovak Republic', 'Syrian Arab Republic',
 5           'United States', 'Venezuela, RB', 'Yemen, Rep.'
 6          ]
 7
 8country_map = {}
 9for c,p in zip(unmatched_covid_countries, map_pop):
10    country_map[p] = c
11
12pop_2019_df['Covid_Country'] = pop_2019_df.apply(
13    lambda x: country_map[x['Country Name']]
14        if x['Country Name'] in country_map
15        else x['Country Name'],
16    axis = 1)
17
18pop_2019_df.iloc[[1,2,3,21,29,249], :]
19"""
20          Country Name         2019 Covid_Country
211          Afghanistan   38041754.0   Afghanistan
222               Angola   31825295.0        Angola
233              Albania    2854191.0       Albania
2421        Bahamas, The     389482.0       Bahamas
2529   Brunei Darussalam     433285.0        Brunei
26249      United States  328239523.0            US
27"""

Confirm that there are no unexpected countries in covid data that are not in population data.

 1list(confirmed_df[
 2    ~confirmed_df.index
 3    .isin(list(pop_2019_df['Covid_Country']))
 4].index)
 5"""
 6['Diamond Princess', 'Holy See', 'MS Zaandam', 'Taiwan*']
 7"""
 8
 9list(deaths_df[
10    ~deaths_df.index
11    .isin(list(pop_2019_df['Covid_Country']))
12].index)
13"""
14['Diamond Princess', 'Holy See', 'MS Zaandam', 'Taiwan*']
15"""

Calculate cases and deaths per 100,000 of population

Merge the population numbers for 2019 into the COVID datasets. Rename the column name for 2019 to Population. This is done for the Confirmed cases and the Deaths datasets.

 1# Merge Population data
 2confirmed_pop_df = confirmed_df.merge(
 3    pop_2019_df,
 4    left_index = True,
 5    left_on = confirmed_df.index,
 6    right_on = 'Covid_Country',
 7)
 8
 9# Rename '2019' column name to 'population'
10confirmed_pop_df = confirmed_pop_df.rename(columns={'2019': 'Population'})
11
12confirmed_pop_df.shape
13"""(187, 329)
14"""
15
16confirmed_pop_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-4,-3,-2,-1]]
17"""
18(187, 329)
19     2020-01-22 00:00:00  2020-01-23 00:00:00  2020-01-24 00:00:00  2020-12-12 00:00:00 Country Name  Population Covid_Country  
201                      0                    0                    0                48229  Afghanistan  38041754.0   Afghanistan  
213                      0                    0                    0                47742      Albania   2854191.0       Albania  
2258                     0                    0                    0                91638      Algeria  43053054.0       Algeria  
23260                    0                    0                    0                 2083  Yemen, Rep.  29161922.0         Yemen  
24262                    0                    0                    0                18217       Zambia  17861030.0        Zambia  
25263                    0                    0                    0                11219     Zimbabwe  14645468.0      Zimbabwe  
26"""

 1# Merge Population data
 2deaths_pop_df = deaths_df.merge(
 3    pop_2019_df,
 4    left_index = True,
 5    left_on = deaths_df.index,
 6    right_on = 'Covid_Country',
 7)
 8
 9# Rename '2019' column name to 'population'
10deaths_pop_df = deaths_pop_df.rename(columns={'2019': 'Population'})
11
12deaths_pop_df.shape
13"""(187, 329)
14"""
15
16deaths_pop_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-4,-3,-2,-1]]
17"""
18     2020-01-22 00:00:00  2020-01-23 00:00:00  2020-01-24 00:00:00  2020-12-12 00:00:00 Country Name  Population Covid_Country  
191                      0                    0                    0                 1956  Afghanistan  38041754.0   Afghanistan  
203                      0                    0                    0                  989      Albania   2854191.0       Albania  
2158                     0                    0                    0                 2584      Algeria  43053054.0       Algeria  
22260                    0                    0                    0                  606  Yemen, Rep.  29161922.0         Yemen  
23262                    0                    0                    0                  366       Zambia  17861030.0        Zambia  
24263                    0                    0                    0                  307     Zimbabwe  14645468.0      Zimbabwe  
25"""

To adjust the numbers of confirmed cases and deaths to 100,000 of the population, divide the number by the population and multiply by 100,000.

 1# Confirmed cases per 100,000 of population
 2confirmed_per_100k_df = confirmed_pop_df.copy()
 3col_names = list(confirmed_per_100k_df.columns)
 4[(col_names.remove(x)) for x in ['Country Name', 'Population', 'Covid_Country']]
 5confirmed_per_100k_df[col_names] = (
 6    confirmed_per_100k_df[col_names]
 7    .div(confirmed_per_100k_df['Population'], axis = 0)
 8    .mul(100000, axis = 0)
 9)
10
11confirmed_per_100k_df.shape
12"""
13(187, 329)
14"""
15
16confirmed_per_100k_df.iloc[[0,1,2,-3,-2,-1], [0,50,100,150,-4,-3,-2,-1]]
17"""
18     2020-01-22 00:00:00  2020-03-12 00:00:00  2020-05-01 00:00:00  2020-06-20 00:00:00  2020-12-12 00:00:00 Country Name  Population  Covid_Country  
191                    0.0             0.031544             6.022330            74.728416           126.779117  Afghanistan  38041754.0    Afghanistan  
203                    0.0             0.805833            27.398307            66.253450          1672.698148      Albania   2854191.0        Albania  
2158                   0.0             0.055745             9.648561            27.015505           212.849012      Algeria  43053054.0        Algeria  
22260                  0.0             0.000000             0.024004             3.161657             7.142876  Yemen, Rep.  29161922.0          Yemen  
23262                  0.0             0.000000             0.610267             8.006257           101.992998       Zambia  17861030.0         Zambia  
24263                  0.0             0.000000             0.273122             3.270636            76.603902     Zimbabwe  14645468.0       Zimbabwe  
25"""

1# Deaths per 100,000 of population
2deaths_per_100k_df = deaths_pop_df.copy()
3col_names = list(deaths_per_100k_df.columns)
4[(col_names.remove(x)) for x in ['Country Name', 'Population', 'Covid_Country']]
5deaths_per_100k_df[col_names] = (
6    deaths_per_100k_df[col_names]
7    .div(deaths_per_100k_df['Population'], axis = 0)
8    .mul(100000, axis = 0)
9)

Countries with highest confirmed cases per 100,000 of population

Countries with populations of less than one million were removed from the dataset. The data is sorted by the numbers of confirmed cases per 100,000 on the latest date and the data for the top 20 countries displayed in a table and bar chart. The changes over time for the top 10 countries is displayed in a line chart.

 1confirmed_per_100k_df = confirmed_per_100k_df[confirmed_per_100k_df['Population'] > 1000000]
 2
 3top_20_confirmed_df = (confirmed_per_100k_df.sort_values(
 4    by = confirmed_per_100k_df.columns[-4], ascending = False)
 5    .head(20)[['Covid_Country',
 6               'Population',
 7               pd.to_datetime('2020-12-12 00:00:00') ]])
 8
 9# Merge Confirmed data
10top_20_confirmed_df = top_20_confirmed_df.merge(
11    confirmed_df.iloc[:, [-1]],
12    right_index = True,
13    left_on = top_20_confirmed_df['Covid_Country'],
14    right_on = confirmed_df.index,
15)
16
17# Reset the index
18top_20_confirmed_df.reset_index(drop = True, inplace = True)
19
20# Drop the merged key
21top_20_confirmed_df = top_20_confirmed_df.drop(['key_0'], axis = 1)
22
23# Rename the column names
24top_20_confirmed_df = top_20_confirmed_df.rename(
25    columns = {'Covid_Country': 'Country',
26               '2020-12-12 00:00:00_x': 'Cases per 100k',
27               '2020-12-12 00:00:00_y': 'Total cases'})

Table: The top 20 countries with highest cases of COVID-19 on November 12 2020

Rank	Country	Population	Cases per 100k	Total cases
1	Bahrain	1,641,172	5,421	88,965
2	Czechia	10,669,709	5,393	575,422
3	Belgium	11,484,055	5,252	603,159
4	Georgia	3,720,382	5,027	187,006
5	Armenia	2,957,731	4,981	147,312
6	Qatar	2,832,067	4,973	140,827
7	US	328,239,523	4,893	16,062,299
8	Moldova	2,657,637	4,731	125,723
9	Slovenia	2,087,946	4,573	95,481
10	Panama	4,246,439	4,488	190,585
11	Switzerland	8,574,832	4,360	373,831
12	Croatia	4,067,500	4,241	172,523
13	Israel	9,053,300	3,930	355,786
14	Serbia	6,944,975	3,764	261,437
15	Spain	47,076,781	3,676	1,730,575
16	Austria	8,877,067	3,603	319,822
17	France	67,059,887	3,587	2,405,255
18	Netherlands	17,332,850	3,540	613,630
19	North Macedonia	2,083,459	3,505	73,025
20	Kuwait	4,207,083	3,471	146,044

 1top_10 = top_20_confirmed_df.head(10)
 2bg_color = 'rgba(208, 225, 242, 1.0)'
 3
 4fig = go.Figure(
 5    data = [go.Bar(x = list(top_10['Cases per 100k']),
 6                   y = list(top_10['Country']),
 7                   hovertemplate = [f"""<b>{r['Country']}</b>: Confirmed Cases<br>
 8<br>Per 100k: {r['Cases per 100k']:,.0F}
 9<br>Total: {r['Total cases']:,.0f}
10<br>Population: {r['Population']:,.0f}
11<extra></extra>
12""" for i,r in top_10.iterrows()],
13                   orientation = 'h')]
14)
15
16fig.update_layout(
17    # Set default font
18        font = dict(
19        family = "Droid Sans",
20        size = 16
21    ),
22    # Set figure title
23    title = dict(
24        text = f"""<b>Countries with highest confirmed cases of COVID-19 per 100,000
25<BR>{pd.to_datetime('2020-12-12 00:00:00').strftime('%b %d %Y')}</b>""",
26        xref = 'container',
27        yref = 'container',
28        x = 0.5,
29        y = 0.91,
30        xanchor = 'center',
31        yanchor = 'middle',
32        font = dict(family = 'Droid Sans', size = 24)
33    ),
34    # set x-axis
35    xaxis = dict(
36        title = dict(
37            text = 'Number of confirmed cases per 100,000 of population',
38            font = dict(family = 'Droid Sans', size = 18)
39        ),
40        showgrid = False,
41        linecolor = bg_color,
42        linewidth = 2,
43        showticklabels = True,
44    ),
45    # set y-axis
46    yaxis = dict(
47        showgrid = False,
48        linecolor = bg_color,
49        linewidth = 4,
50    ),
51    # set the plot bacground color
52    plot_bgcolor = bg_color,
53    # set the hover background color
54    hoverlabel = dict(
55        bgcolor = 'rgba(75, 152, 201, 0.2)',
56        font_size = 16
57    ),
58    paper_bgcolor = bg_color,
59)
60
61fig.update_traces(marker_color = 'rgb(75, 152, 201)')
62fig['layout']['yaxis']['autorange'] = "reversed"
63
64fig.show()

Highest confirmed cases of COVID-19 per 100,000 of population

 1top_10_confirmed_df = (confirmed_per_100k_df.sort_values(
 2    by = confirmed_per_100k_df.columns[-4], ascending = False)
 3    .head(10))
 4
 5top_10 = (top_10_confirmed_df
 6          .drop(['Country Name', 'Population'], axis = 1)
 7          .set_index('Covid_Country')
 8          .T)
 9
10bg_color = 'rgba(208, 225, 242, 1.0)'
11line_color = 'rgba(75, 152, 201, 1.0)'
12grid_color = 'rgba(75, 152, 201, 0.3)'
13
14fig = go.Figure()
15for c in top_10.columns:
16    fig.add_trace(go.Scatter(
17        x = top_10.index,
18        y = top_10[c],
19        name = c,
20        hovertemplate = [f"""<b>{c}</b><br>
21<br>Cases per 100k:<br>{r[c]:,.0F}
22<extra></extra>
23""" for i,r in top_10.iterrows()]
24    ))
25
26fig.update_traces(
27     hoverinfo = 'text+name',
28    mode = 'lines'
29)
30
31fig.update_layout(
32    # Set figure title
33    title = dict(
34        text = f"""<b>Highest confirmed cases of COVID-19 per 100,000</b>\
35<br>{top_10.index[-1].strftime('%b %d %Y')}""",
36        xref = 'container',
37        yref = 'container',
38        x = 0.5,
39        y = 0.9,
40        xanchor = 'center',
41        yanchor = 'middle',
42        font = dict(family = 'Droid Sans', size = 28)
43    ),
44    # set legend
45    legend = dict(
46        orientation = 'v',
47        traceorder = 'normal',
48        font_size = 12,
49        x = 0.05,
50        y = 1.0,
51        xanchor = "left",
52        yanchor = "top"
53    ),
54    # set x-axis
55    xaxis = dict(
56        title = 'Date',
57        linecolor = line_color,
58        linewidth = 2,
59        gridcolor = grid_color,
60        showticklabels = True,
61        ticks = 'outside',
62    ),
63    # set y-axis
64    yaxis = dict(
65        title = 'Cases of COVID-19 per 100.000 of population',
66        rangemode = 'tozero',
67        linecolor = line_color,
68        linewidth = 2,
69        gridcolor = grid_color,
70        showticklabels = True,
71        ticks = 'outside',
72    ),
73    # set the plot background color
74    plot_bgcolor = bg_color,
75    paper_bgcolor = bg_color,
76)
77
78fig.show()

Changes in countries with highest confirmed cases of COVID-19 per 100,000 of population

Countries with highest deaths per 100,000 of population

Similar table and charts are created for the Deaths from COVID-19 data.

Table: The top 20 countries with highest deaths per 100,000 from COVID-19 on November 12 2020

Rank	Country	Population	Deaths per 100k	Total deaths
1	Belgium	11,484,055	154.9	17,792
2	Peru	32,510,453	112.4	36,544
3	Italy	60,297,396	106.2	64,036
4	Spain	47,076,781	101.2	47,624
5	North Macedonia	2,083,459	100.6	2,096
6	Bosnia and Herzegovina	3,301,000	99.9	3,298
7	Slovenia	2,087,946	97.8	2,041
8	United Kingdom	66,834,405	95.9	64,123
9	Moldova	2,657,637	95.8	2,547
10	US	328,239,523	90.7	297,818
11	Argentina	44,938,712	90.5	40,668
12	Mexico	127,575,529	89.1	113,704
13	Czechia	10,669,709	88.6	9,450
14	France	67,059,887	86.0	57,671
15	Brazil	211,049,527	85.8	181,123
16	Chile	18,952,038	83.6	15,846
17	Armenia	2,957,731	83.2	2,462
18	Bulgaria	6,975,761	80.7	5,626
19	Ecuador	17,373,662	79.9	13,874
20	Panama	4,246,439	78.4	3,331

Highest deaths from COVID-19 per 100,000 of population

Changes in countries with highest deaths from COVID-19 per 100,000 of population

Conclusion

Calculations can be applied to a whole dataset easily with Pandas, such as converting the total numbers of COVID-19 cases to numbers relative to a standard number. The challenge here was getting the population, which is available from the World Data Bank, but the names for countries are inconsistent. Once these are mapped to the countries, the population is merged in for each country. The div and mul dataframe functions are used to adjust the numbers relative to 100,000 of the population. The countries with highest relative numbers are not necessarily the same countries with the highest absolute numbers.

Plotly is an open-source graphing library for Python that produces interactive charts.

Add data with a secondary y axis

Eric Callanan — Sunday, 06 Dec 2020

Display two sets of data on the same chart when the data ranges are different, such as the confirmed cases of COVID-19 and deaths from COVID-19. COVID-19 is the disease caused by a new coronavirus (SARS-CoV-2) that the World Health Organisation (WHO) declared a pandemic in March 2020. This article will show how to create an interactive line chart with plotly, showing the changes in confirmed cases and deaths over time.

The data used in this article is retrieved from the Johns Hopkins University who have made the data available on GitHub. More information about COVID-19 and the coronavirus is available from Coronavirus disease (COVID-19) advice for the public. Plotly is an open-source graphing library for Python that produces interactive charts.

Load the data into a Pandas dataframe

The data is available on John Hopkins GitHub page. There are two sets of data to load the global confirmed cases and the global deaths.

time_series_covid19_deaths_global.csv
time_series_covid19_confirmed_global.csv

The data can be loaded directly from the GitHub page as in the code below. The data is cleaned to remove unwanted fields and group the data by country with the following function.

 1def load_clean_data(csv_path):
 2    df = pd.read_csv(csv_path)
 3
 4    # 1. Drop unwanted columns
 5    df.drop(['Province/State', 'Lat', 'Long'], axis = 1, inplace = True)
 6
 7    # 2. Group by Country
 8    df = df.groupby('Country/Region').sum()
 9
10    # 3. Transpose data to put dates in a single column
11    df = df.T
12    df.index = pd.to_datetime(df.index)
13
14    # 4. Remove the column name
15    df.rename_axis(None, axis = 1, inplace = True)
16
17    return df

Load the data for Confirmed cases and the data for Deaths from COVID-19.

 1
 2confirmed_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv')
 3confirmed_df.shape
 4"""
 5(314, 191)
 6"""
 7
 8confirmed_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
 9"""
10            Afghanistan  Albania  Algeria  Yemen  Zambia  Zimbabwe
112020-01-22            0        0        0      0       0         0
122020-01-23            0        0        0      0       0         0
132020-01-24            0        0        0      0       0         0
142020-11-28        45966    36790    81212   2160   17589      9822
152020-11-29        46215    37625    82221   2177   17608      9822
162020-11-30        46498    38182    83199   2191   17647      9950
17"""
18
19
20deaths_df = load_clean_data('https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv')
21deaths_df.shape
22"""
23(314, 191)
24"""
25
26deaths_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
27"""
28            Afghanistan  Albania  Algeria  Yemen  Zambia  Zimbabwe
292020-01-22            0        0        0      0       0         0
302020-01-23            0        0        0      0       0         0
312020-01-24            0        0        0      0       0         0
322020-11-28         1752      787     2393    615     357       275
332020-11-29         1763      798     2410    617     357       275
342020-11-30         1774      810     2431    619     357       276
35"""

Global confirmed cases of COVID-19

Show the total global number of confirmed cases of COVID-19 as the numbers changes over time. Plotly is used to display an interactive line chart. An annotation is added to show the current total number on the latest date.

 1def hov_text(ary_x, ary_y, name):
 2    '''Create the hover text for each data point
 3
 4    Keyword arguments:
 5    ary_x    -- list of all the x data points
 6    ary_y    -- list of all the y data points
 7    name     -- name to display before each value
 8
 9    return: A list of hover texts for the selected data
10            with format to display
11    '''
12    txt = [f'''
13<b>{ary_x[i].strftime('%B %d %Y')}</b><br><br>
14{name}: <b>{ary_y[i]:,.0F}</b><br>
15<extra></extra>
16''' for i in range(len(ary_x))]
17    return txt

 1total_confirmed = confirmed_df.sum(axis = 1)
 2
 3bg_color = 'rgba(208, 225, 242, 1.0)'
 4line_color = 'rgba(75, 152, 201, 1.0)'
 5grid_color = 'rgba(75, 152, 201, 0.3)'
 6
 7fig = go.Figure()
 8fig.add_trace(go.Scatter(x = total_confirmed.index,
 9                         y = total_confirmed,
10                         name = 'Confirmed',
11                         hovertemplate = hov_text(total_confirmed.index,
12                                                  total_confirmed,
13                                                  'Confirmed Cases')))
14
15fig.update_traces(
16    hoverinfo = 'text+name',
17    mode = 'lines'
18)
19
20fig.update_layout(
21    # Set figure title
22    title = dict(
23        text = f'<b>Global confirmed cases from COVID-19</b>',
24        xref = 'container',
25        yref = 'container',
26        x = 0.5,
27        y = 0.9,
28        xanchor = 'center',
29        yanchor = 'middle',
30        font = dict(family = 'Droid Sans', size = 28)
31    ),
32    # set legend
33    legend = dict(
34        orientation = 'h',
35        traceorder = 'normal',
36        font_size = 12,
37        x = 0.0,
38        y = -0.3,
39        xanchor = 'left',
40        yanchor = 'top'
41    ),
42    # set x-axis
43    xaxis = dict(
44        title = 'Date',
45        linecolor = line_color,
46        linewidth = 2,
47        gridcolor = grid_color,
48        showticklabels = True,
49        ticks = 'outside',
50    ),
51    # set y-axis
52    yaxis = dict(
53        title = 'Number of confirmed cases',
54        rangemode = 'tozero',
55        linecolor = line_color,
56        linewidth = 2,
57        gridcolor = grid_color,
58        showticklabels = True,
59        ticks = 'outside',
60    ),
61    showlegend = False,
62    # set the plot bacground color
63    plot_bgcolor = bg_color,
64    paper_bgcolor = bg_color,
65)
66
67# Add text for current date and total
68fig.add_annotation(
69    text = f"{total_confirmed.index[-1].strftime('%B %d')}<BR> {total_confirmed[-1]:,.0F}",
70    xref = 'paper',
71    yref = 'paper',
72    x = 0.5,
73    y = 1.0,
74    showarrow = False,
75    bgcolor = 'rgba(180, 210, 233, 1.0)',
76    borderpad = 10,
77    font = dict(family = 'Droid Sans', size = 28)
78)
79
80fig.show()

Global confirmed cases from COVID-19

Global deaths from COVID-19

Show the total global number of deaths from COVID-19 as the numbers changes over time. The code is similar to the last chart except the data source and titles.

 1total_deaths = deaths_df.sum(axis = 1)
 2
 3bg_color = 'rgba(208, 225, 242, 1.0)'
 4line_color = 'rgba(75, 152, 201, 1.0)'
 5grid_color = 'rgba(75, 152, 201, 0.3)'
 6
 7fig = go.Figure()
 8fig.add_trace(go.Scatter(x = total_deaths.index,
 9                         y = total_deaths,
10                         name = 'Deaths',
11                         hovertemplate = hov_text(total_deaths.index,
12                                                  total_deaths,
13                                                 'Deaths')))
14
15fig.update_traces(
16    hoverinfo = 'text+name',
17    mode = 'lines'
18)
19
20fig.update_layout(
21    # Set figure title
22    title = dict(
23        text = f'<b>Global deaths from COVID-19</b>',
24        xref = 'container',
25        yref = 'container',
26        x = 0.5,
27        y = 0.9,
28        xanchor = 'center',
29        yanchor = 'middle',
30        font = dict(family = 'Droid Sans', size = 28)
31    ),
32    # set legend
33    legend = dict(
34        orientation = 'h',
35        traceorder = 'normal',
36        font_size = 12,
37        x = 0.0,
38        y = -0.3,
39        xanchor = 'left',
40        yanchor = 'top'
41    ),
42    # set x-axis
43    xaxis = dict(
44        title = 'Date',
45        linecolor = line_color,
46        linewidth = 2,
47        gridcolor = grid_color,
48        showticklabels = True,
49        ticks = 'outside',
50    ),
51    # set y-axis
52    yaxis = dict(
53        title = 'Number of deaths',
54        rangemode = 'tozero',
55        linecolor = line_color,
56        linewidth = 2,
57        gridcolor = grid_color,
58        showticklabels = True,
59        ticks = 'outside',
60    ),
61    showlegend = False,
62    # set the plot bacground color
63    plot_bgcolor = bg_color,
64    paper_bgcolor = bg_color,
65)
66
67# Add text for current date and total
68fig.add_annotation(
69    text = f"{total_deaths.index[-1].strftime('%B %d')}<BR> {total_deaths[-1]:,.0F}",
70    xref = 'paper',
71    yref = 'paper',
72    x = 0.5,
73    y = 1.0,
74    showarrow = False,
75    bgcolor = 'rgba(180, 210, 233, 1.0)',
76    borderpad = 10,
77    font = dict(family = 'Droid Sans', size = 28)
78)
79
80fig.show()

Global deaths from COVID-19

Show Confirmed cases and deaths from COVID-19 on a single graph

Just showing the data for confirmed cases and deaths on the same chart does not work well as there is such a difference in scale between the number ranges. The number of deaths looks flat relative to the total number of cases, but these numbers are growing.

Global deaths from COVID-19

One solution is to plot the Deaths data using a second y axis so that each set of data will use the full vertical space available. These y axes are also marked with color associated with the data that uses the axis.

The code to generate this chart is wrapped up in a function that takes the two dataframes and the chart title and returns a plotly graph figure.

 1def plot_confirmed_and_deaths(df1, df2, region):
 2    bg_color = 'rgba(208, 225, 242, 1.0)'
 3    line_color = 'rgba(75, 152, 201, 1.0)'
 4    grid_color = 'rgba(75, 152, 201, 0.3)'
 5
 6    # Linechart with secondary y axis
 7    fig = make_subplots(specs=[[{"secondary_y": True}]])
 8
 9    fig.add_trace(
10        go.Scatter(
11            x = df1.index,
12            y = df1,
13            name = 'Confirmed Cases',
14            hovertemplate = hov_text(df1.index,
15                                     df1,
16                                     'Confirmed Cases')),
17        secondary_y = False)
18
19    fig.add_trace(
20        go.Scatter(
21            x = df2.index,
22            y = df2,
23            name = 'Deaths',
24            line = dict(dash = 'dashdot'),
25            hovertemplate = hov_text(df2.index,
26                                     df2,
27                                     'Deaths')),
28        secondary_y = True)
29
30    fig.update_layout(
31        # Set figure title
32        title = dict(
33            text = f'Confirmed cases and deaths from COVID-19<BR><b>{region}</b>',
34            xref = 'container',
35            yref = 'container',
36            x = 0.5,
37            y = 0.9,
38            xanchor = 'center',
39            yanchor = 'middle',
40            font = dict(family = 'Droid Sans', size = 28)
41        ),
42        # set legend
43        legend = dict(
44            orientation = 'v',
45            traceorder = 'normal',
46            font_size = 12,
47            x = 0.1,
48            y = 0.9,
49            xanchor = "left",
50            yanchor = "top"
51        ),
52        # set x-axis
53        xaxis = dict(
54            title = 'Date',
55            linecolor = line_color,
56            linewidth = 2,
57            gridcolor = grid_color,
58            showticklabels = True,
59            ticks = 'outside',
60        ),
61        # set y-axis
62        yaxis = dict(
63            title = 'Number of Confirmed Cases',
64            color = 'rgba(80, 80, 250, 1.0)',
65            rangemode = 'tozero',
66            linecolor = 'rgba(80, 80, 250, 1.0)',
67            linewidth = 2,
68            gridcolor = 'rgba(80, 80, 250, 0.3)',
69            showticklabels = True,
70            ticks = 'outside'
71        ),
72        # set the plot bacground color
73        plot_bgcolor = bg_color,
74        paper_bgcolor = bg_color,
75    )
76
77    # set the secondary y-axis
78    fig.update_yaxes(
79        title_text = 'Number of Deaths',
80        color = 'rgba(237, 34, 13, 1.0)',
81        range = [0, (df2[-1] * 1.08)],
82        linecolor = 'rgba(237, 34, 13, 1.0)',
83        linewidth = 2,
84        gridcolor = 'rgba(237, 34, 13, 0.3)',
85        showticklabels = True,
86        ticks = 'outside',
87        secondary_y = True
88    )
89
90    return fig

This plot_confirmed_and_deaths function is used to create the line plot showing both confirmed cases and deaths from COVID-19 globally.

1df1 = confirmed_df.sum(axis = 1)
2df2 = deaths_df.sum(axis = 1)
3
4f = plot_confirmed_and_deaths(df1, df2, 'Global')
5f.show()

Global deaths from COVID-19

Display data for selected country

The same function can be used to create a chart for any country.

1country = 'US'
2f = plot_confirmed_and_deaths(
3    confirmed_df[country],
4    deaths_df[country],
5    country)
6f.show()

Global deaths from COVID-19 in United States

Global deaths from COVID-19 in United Kingdom

Global deaths from COVID-19 in Ireland

Global deaths from COVID-19 in Brazil

Conclusion

Pandas dataframe is great for loading data from csv files and filtering the data to an area of interest quickly. Plotly is great to create interactive charts. It is informative to see the rise in the number of deaths from COVID-19 in relation to the rise in the total number of cases detected. This relationship is easier to see when the data can be shown on a single graph. A secondary y-axis is used to display the deaths as the scales are so different. All of the loading, cleaning and creation of the chart can be wrapped up in a couple of functions to create a pipeline to generate these charts for countries of interest. The interactive charts allows the viewer to zoom in on a particular date range and select or deselect data to focus on an area of interest.

Countries with highest confirmed cases of Covid-19 with Plotly

Eric Callanan — Sunday, 15 Nov 2020

COVID-19 is the disease caused by a new coronavirus called SARS-CoV-2. World Health Organisation (WHO) first learned of the virus on 31 December 2019. The WHO declared the coronavirus outbreak a pandemic in March 2020. This article will show how to create an interactive line chart with plotly, showing the changes in confirmed cases over time.

The data used in this article is retrieved from the Johns Hopkins University who have made te data available on GitHub. More information about COVID-19 and the coronavirus is available from Coronavirus disease (COVID-19) advice for the public.

Plotly is an open-source graphing library for Python that produces interactive charts.

Load the data into a Pandas dataframe

The data is available on John Hopkins GitHub page. The data for the daily deaths from corona virus is in time_series_covid19_deaths_global.csv file. This file can either be downloaded and loaded into a dataframe or it can be loaded directly from GitHub as in the code below. Load and review the data.

 1confirmed_path = 'https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv'
 2confirmed_df = pd.read_csv(confirmed_path)
 3confirmed_df.shape
 4"""
 5(271, 315)
 6"""
 7
 8confirmed_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
 9"""
10    Province/State Country/Region        Lat  11/25/20  11/26/20  11/27/20
110              NaN    Afghanistan  33.939110     45490     45716     45839
121              NaN        Albania  41.153300     34944     35600     36245
132              NaN        Algeria  28.033900     78025     79110     80168
14268            NaN          Yemen  15.552727      2124      2137      2148
15269            NaN         Zambia -13.133897     17535     17553     17569
16270            NaN       Zimbabwe -19.015438      9508      9623      9714
17"""

The data in the Country/Region is further broken down by Province/State for some countries such as China or United Kingdom. A single row is created for each country by grouping the data by Country/Region and summing up the number. the columns 'Province/State', 'Lat', 'Long' are removed from the dataset.

 1# Drop columns 'Province/State', 'Lat', 'Long'
 2confirmed_df.drop(['Province/State', 'Lat', 'Long'], axis = 1, inplace = True)
 3
 4country_df = confirmed_df.groupby('Country/Region').sum()
 5country_df.shape
 6"""
 7(271, 312)
 8"""
 9
10country_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
11"""
12    Country/Region  1/22/20  1/23/20  11/25/20  11/26/20  11/27/20
130      Afghanistan        0        0     45490     45716     45839
141          Albania        0        0     34944     35600     36245
152          Algeria        0        0     78025     79110     80168
16268          Yemen        0        0      2124      2137      2148
17269         Zambia        0        0     17535     17553     17569
18270       Zimbabwe        0        0      9508      9623      9714
19"""

Display confirmed cases in top ten countries on the latest date

Sort the data by the values in the last column in descending order. A bar chart is created using Plotly with the top ten values. This displays an interactive chart that displays the value for each country when the bar is hovered over.

 1Top_10 = (country_df
 2          .sort_values(by=country_df.columns[-1], ascending=False)
 3          [[country_df.columns[-1]]]
 4          .head(10))
 5bg_color = 'rgba(208, 225, 242, 1.0)'
 6
 7fig = go.Figure(
 8    data = [go.Bar(x = list(Top_10.iloc[:, 0]),
 9                   y = list(Top_10.index),
10                   hovertemplate = "%{y}: <br><br>Confirmed: %{x:,.0f}<extra></extra>",
11                   orientation = 'h')]
12)
13
14fig.update_layout(
15    # Set default font
16        font = dict(
17        family = "Droid Sans",
18        size = 16
19    ),
20    # Set figure title
21    title = dict(
22        text = "<b>Top ten countries with highest confirmed cases of covid-19</b>",
23        xref = 'container',
24        yref = 'container',
25        x = 0.5,
26        y = 0.91,
27        xanchor = 'center',
28        yanchor = 'middle',
29        font = dict(family = 'Droid Sans', size = 24)
30    ),
31    # set x-axis
32    xaxis = dict(
33        title = dict(
34            text = 'Number of confirmed cases',
35            font = dict(family = 'Droid Sans', size = 18)
36        ),
37        showgrid = False,
38        linecolor = bg_color,
39        linewidth = 2,
40        showticklabels = True,
41    ),
42    # set y-axis
43    yaxis = dict(
44        showgrid = False,
45        linecolor = bg_color,
46        linewidth = 4,
47    ),
48    # set the plot bacground color
49    plot_bgcolor = bg_color,
50    # set the hover background color
51    hoverlabel = dict(
52        bgcolor = 'rgba(75, 152, 201, 0.2)',
53        font_size = 16
54    ),
55    paper_bgcolor = bg_color,
56)
57
58# Add annotation for the latest date
59fig.add_annotation(
60    xref = "paper",
61    yref = "paper",
62    x = 0.9,
63    y = 0.2,
64    text = f"<b>{pd.to_datetime(Top_10.columns[0]).strftime('%b %d %Y')}</b>",
65    showarrow = False,
66    font = dict(family = 'Droid Sans', size = 40))
67
68fig.update_traces(marker_color = 'rgb(75, 152, 201)')
69fig['layout']['yaxis']['autorange'] = "reversed"
70
71fig.show()

Top ten countries with highest confirmed cases of COVID-19

Display confirmed cases in US

Display a line chart showing the changes in the number of confirmed case over time for a specific country. Start with the US as this has highest number of confirmed cases in the world. The index is converted to datetime format so the axis is displayed better.

 1us_data = country_df.loc['US']
 2# Convert the index to datetime format
 3us_data.index = pd.to_datetime(us_data.index)
 4
 5def hov_text(ary_x, ary_y):
 6    '''Create the hover text for each data point
 7
 8    Keyword arguments:
 9    ary_x    -- list of all the x data points
10    ary_y    -- list of all the y data points
11
12    return: A list of hover texts for the selected country
13            with format to display
14    '''
15    txt = [f'''
16<b>{ary_x[i].strftime('%b %d %Y')}</b><br><br>
17Confirmed Cases: <b>{ary_y[i]:,.0F}</b><br>
18<extra></extra>
19''' for i in range(len(ary_x))]
20    return txt
21
22bg_color = 'rgba(208, 225, 242, 1.0)'
23line_color = 'rgba(75, 152, 201, 1.0)'
24grid_color = 'rgba(75, 152, 201, 0.3)'
25
26fig = go.Figure()
27fig.add_trace(go.Scatter(x = us_data.index,
28                         y = us_data,
29                         name = 'US',
30                         hovertemplate = hov_text(us_data.index, us_data)))
31
32fig.update_traces(
33    hoverinfo = 'text+name',
34    mode = 'lines'
35)
36
37fig.update_layout(
38    # Set figure title
39    title = dict(
40        text = f'<b>Confirmed cases of COVID-19 in US</b>',
41        xref = 'container',
42        yref = 'container',
43        x = 0.5,
44        y = 0.9,
45        xanchor = 'center',
46        yanchor = 'middle',
47        font = dict(family = 'Droid Sans', size = 28)
48    ),
49    # set legend
50    legend = dict(
51        orientation = "h",
52        traceorder = 'normal',
53        font_size = 12,
54        x = 0.0,
55        y = -0.3,
56        xanchor = "left",
57        yanchor = "top"
58    ),
59    # set x-axis
60    xaxis = dict(
61        title = 'Date',
62        linecolor = line_color,
63        linewidth = 2,
64        gridcolor = grid_color,
65        showticklabels = True,
66        ticks = 'outside',
67    ),
68    # set y-axis
69    yaxis = dict(
70        title = 'Number of confirmed cases',
71        linecolor = line_color,
72        linewidth = 2,
73        gridcolor = grid_color,
74        showticklabels = True,
75        ticks = 'outside',
76    ),
77    # set the plot bacground color
78    plot_bgcolor = bg_color,
79    paper_bgcolor = bg_color,
80)
81
82fig.show()

Confirmed cases of COVID-19 in US

Create a function to display confirmed cases for any country

The code above can be wrapped up in a function to create a chart for any country specified.

 1def hov_text(ary_x, ary_y):
 2    '''Create the hover text for each data point
 3
 4    Keyword arguments:
 5    ary_x    -- list of all the x data points
 6    ary_y    -- list of all the y data points
 7
 8    return: A list of hover texts for the selected country
 9            with format to display
10    '''
11    txt = [f'''
12<b>{ary_x[i].strftime('%b %d %Y')}</b><br><br>
13Confirmed Cases: <b>{ary_y[i]:,.0F}</b><br>
14<extra></extra>
15''' for i in range(len(ary_x))]
16    return txt
17
18
19def plot_confirmed_cases_for_country(df, country):
20    sel_data = df.loc[country]
21    # Convert the index to datetime format
22    sel_data.index = pd.to_datetime(sel_data.index)
23
24    bg_color = 'rgba(208, 225, 242, 1.0)'
25    line_color = 'rgba(75, 152, 201, 1.0)'
26    grid_color = 'rgba(75, 152, 201, 0.3)'
27
28    fig = go.Figure()
29    fig.add_trace(go.Scatter(x = sel_data.index,
30                             y = sel_data,
31                             name = 'US',
32                             hovertemplate = hov_text(sel_data.index, sel_data)))
33
34    fig.update_traces(
35        hoverinfo = 'text+name',
36        mode = 'lines'
37    )
38
39    fig.update_layout(
40        # Set figure title
41        title = dict(
42            text = f'<b>Confirmed cases of COVID-19 in {country}</b>',
43            xref = 'container',
44            yref = 'container',
45            x = 0.5,
46            y = 0.9,
47            xanchor = 'center',
48            yanchor = 'middle',
49            font = dict(family = 'Droid Sans', size = 28)
50        ),
51        # set legend
52        legend = dict(
53            orientation = "h",
54            traceorder = 'normal',
55            font_size = 12,
56            x = 0.0,
57            y = -0.3,
58            xanchor = "left",
59            yanchor = "top"
60        ),
61        # set x-axis
62        xaxis = dict(
63            title = 'Date',
64            linecolor = line_color,
65            linewidth = 2,
66            gridcolor = grid_color,
67            showticklabels = True,
68            ticks = 'outside',
69        ),
70        # set y-axis
71        yaxis = dict(
72            title = 'Number of confirmed cases',
73            linecolor = line_color,
74            linewidth = 2,
75            gridcolor = grid_color,
76            showticklabels = True,
77            ticks = 'outside',
78        ),
79        # set the plot bacground color
80        plot_bgcolor = bg_color,
81        paper_bgcolor = bg_color,
82    )
83
84    return fig

Call this function with the dataframe and the country name to create the plot.

1fig = plot_confirmed_cases_for_country(country_df, 'China')
2fig.show()

Confirmed cases of COVID-19 in China

Display confirmed cases for top ten countries

Get the ten country names for the countries with the highest confirmed cases on the latest date.

1top_countries = (country_df
2                 .sort_values(by = country_df.columns[-1], ascending = False)
3                 .head(10)
4                 .index)
5
6"""
7['US',   'India',          'Brazil', 'France',    'Russia',
8'Spain', 'United Kingdom', 'Italy',  'Argentina', 'Colombia']
9"""

Transpose the data so that the date is the index and each row represents all the data for that date. Set the index to be datetime format.

 1wide_df = country_df.T
 2wide_df.index = pd.to_datetime(wide_df.index)
 3
 4wide_df.shape
 5"""
 6(311, 191)
 7"""
 8
 9wide_df.iloc[[0,1,2,-3,-2,-1], [0,1,2,-3,-2,-1]]
10"""
11Country/Region  Afghanistan  Albania  Algeria  Yemen  Zambia  Zimbabwe
122020-01-22                0        0        0      0       0         0
132020-01-23                0        0        0      0       0         0
142020-01-24                0        0        0      0       0         0
152020-11-25            45490    34944    78025   2124   17535      9508
162020-11-26            45716    35600    79110   2137   17553      9623
172020-11-27            45839    36245    80168   2148   17569      9714
18"""

The data is extracted for the top ten countries and displayed on a line chart.

 1def hov_text(ary_x, ary_y, country):
 2    '''Create the hover text for each data point
 3
 4    Keyword arguments:
 5    ary_x    -- list of all the x data points
 6    ary_y    -- list of all the y data points
 7    country  -- name of country
 8
 9    return: A list of hover texts for the selected country
10            with format to display
11    '''
12    txt = [f'''
13<b>{country}<b><br><br>
14Date: <b>{ary_x[i].strftime('%b %d %Y')}</b><br>
15Confirmed Cases: <b>{ary_y[i]:,.0F}</b><br>
16<extra></extra>
17''' for i in range(len(ary_x))]
18    return txt
19
20bg_color = 'rgba(208, 225, 242, 1.0)'
21line_color = 'rgba(75, 152, 201, 1.0)'
22grid_color = 'rgba(75, 152, 201, 0.3)'
23
24df = wide_df[top_countries]
25
26fig = go.Figure()
27for c in top_countries:
28    fig.add_trace(go.Scatter(x = df.index,
29                             y = df[c],
30                             name = c,
31                             hovertemplate = hov_text(df.index, df[c], c)))
32
33fig.update_traces(
34    hoverinfo = 'text+name',
35    mode = 'lines'
36)
37
38fig.update_layout(
39    # Set figure title
40    title = dict(
41        text = f'<b>Confirmed cases of COVID-19 in top ten countries</b>',
42        xref = 'container',
43        yref = 'container',
44        x = 0.5,
45        y = 0.9,
46        xanchor = 'center',
47        yanchor = 'middle',
48        font = dict(family = 'Droid Sans', size = 28)
49    ),
50    # set legend
51    legend = dict(
52        orientation = "h",
53        traceorder = 'normal',
54        font_size = 12,
55        x = 0.0,
56        y = -0.3,
57        xanchor = "left",
58        yanchor = "top"
59    ),
60    # set x-axis
61    xaxis = dict(
62        title = 'Date',
63        linecolor = line_color,
64        linewidth = 2,
65        gridcolor = grid_color,
66        showticklabels = True,
67        ticks = 'outside',
68    ),
69    # set y-axis
70    yaxis = dict(
71        title = 'Number of confirmed cases',
72        linecolor = line_color,
73        linewidth = 2,
74        gridcolor = grid_color,
75        showticklabels = True,
76        ticks = 'outside',
77    ),
78    showlegend = True,
79    # set the plot bacground color
80    plot_bgcolor = bg_color,
81    paper_bgcolor = bg_color,
82)
83
84fig.show()

Confirmed cases of COVID-19 for top ten countries

Conclusion

Pandas and Plotly are a powerful combination to load data, clean data and create interactive charts. Johns Hopkins University has done the hard work of collecting the COVID-19 data and making it available. The data can be loaded directly from the csv file in GitHub. Plotly is used to create bar charts and line charts to display information on countries with the highest confirmed cases of COVID-19.

Plotly - Customise plotly toolbar

Eric Callanan — Sunday, 11 Oct 2020

Plotly is an open-source graphing library for Python that is used to create interactive chart. Each chart displays a toolbar on the top right of the chart that contains a set of buttons to interact with the chart. These buttons provide functionality such as download an image of the chart, zoom in and out, pan, reset the axes as well as a logo with a link to Plotly.

The plotly toolbar is called the Modebar and the default is to display when when a cursor is hovered over the chart or the modebar area. The buttons displayed on the modebar vary depending on the type of chart such as 3D charts providing buttons for orbital rotation that do not appear on 2D charts.

All of these buttons can confuse users and sometimes clutter up the chart. It is possible to remove some of the buttons, but the mechanism is not that straight-forward. The configuration of the modebar is associated with the display of the chart and not tied with the creation of the chart itself. In this article a simple line chart is created and the modebar is modified when displaying the chart.

Show or hide the Modebar

The plotly chart can be displayed with the Modebar permanently visible or configured to not show at all.

Create the sample line chart

First create a dataframe with random data.

1samples = 15
2random.seed(42)
3df = pd.DataFrame({
4    'A': sorted([random.randint(1,100) for _ in range(15)]),
5    'B': sorted([random.randint(1,100) for _ in range(15)]),
6    'C': sorted([random.randint(1,100) for _ in range(15)]),
7    'D': sorted([random.randint(1,100) for _ in range(15)]),
8})

Create a line chart with the data in the dataframe.

 1# Create line chart using plotly.graph_objects
 2
 3bg_color = 'rgba(208, 225, 242, 1.0)'
 4line_color = 'rgba(75, 152, 201, 1.0)'
 5grid_color = 'rgba(75, 152, 201, 0.3)'
 6
 7y_max = 110
 8y_min = 0
 9x_min = 0
10x_max = 15
11
12fig = go.Figure()
13for c in df.columns:
14    fig.add_trace(go.Scatter(x = df.index,
15                             y = df[c],
16                             mode='lines',
17                             hoverinfo = 'text+name',
18                             name = c,
19                             hovertemplate = [f'''
20                                <b>{c}<b><br><br>
21                                Value: <b>{df[c][i]:.1F}</b><br>
22                                <extra></extra>
23                                ''' for i in range(len(df.index))]
24                            ))
25
26fig.update_layout(
27    # Set figure title
28    title = dict(
29        text = f'<b>Sample line chart</b>',
30        xref = 'container',
31        yref = 'container',
32        x = 0.5,
33        y = 0.9,
34        xanchor = 'center',
35        yanchor = 'middle',
36        font = dict(family = 'Droid Sans', size = 28)
37    ),
38    # set x-axis
39    xaxis = dict(
40        title = 'x-axis data',
41        range = [x_min, x_max],
42        linecolor = line_color,
43        linewidth = 2,
44        gridcolor = grid_color,
45        showticklabels = True,
46        ticks = 'outside',
47    ),
48    # set y-axis
49    yaxis = dict(
50        title = 'y-axis data',
51        range = [y_min, y_max],
52        linecolor = line_color,
53        linewidth = 2,
54        gridcolor = grid_color,
55        showticklabels = True,
56        ticks = 'outside',
57    ),
58    # set legend
59    legend = dict(
60        traceorder = 'normal',
61        font_size = 12,
62        x = 0.01,
63        y = 0.99,
64        xanchor = "left",
65        yanchor = "top"
66    ),
67    showlegend = True,
68    # set the plot bacground color
69    plot_bgcolor = bg_color,
70    paper_bgcolor = bg_color,
71)
72
73fig.show()

The default is to show the modebar only when a cursor is hovered over the chart or over the modebar.

Simple line chart with default Modebar behavior

Always show the Modebar

This Modebar can be configured to always display using displayModeBar property in a dictionary of values for the config property of the show function.

1fig.show(config={"displayModeBar": True})

Line chart with Modebar always visible

Never show the Modebar

The Modebar can also be configured to never display using displayModeBar property set to false. The three options for displayModeBar are 'hover', True, False. Hiding the Modebar does not change the ability to interact with the chart; hover still works; selecting an area zoomes in on this area and double-click resets the chart.

1fig.show(config={"displayModeBar": False})

Line chart with Modebar never visible

Remove the plotly logo from the Modebar

The Modebar includes the Plotly logo with a link to Plotly at the end of the Modebar. This can be removed with displaylogo parameter.

1fig.show(config={"displaylogo": False})

Line chart without Plotly logo on Modebar

Remove specific buttons from the Modebar

Specific buttons can be removed from the Modebar using the modeBarButtonsToRemove parameters. This takes a list of strings of the names of the buttons to remove from the Modebar. The following is a list of all the modebar buttons, some of which are only present on certain chart types.

Summary of the pandas.concat and dataframe.append:

Chart	Button names
general	toImage, toggleHover, resetViews, sendDataToCloud, toggleSpikelines, resetViewMapbox, hoverClosestGl2d, hoverClosestPie, hoverClosestCartesian, hoverCompareCartesian
2D	zoom2d, pan2d, select2d, lasso2d, zoomIn2d, zoomOut2d, autoScale2d, resetScale2d
3D	zoom3d, pan3d, rbitRotation, tableRotation, handleDrag3d, resetCameraDefault3d, resetCameraLastSave3d, hoverClosest3d
Geo	zoomInGeo, zoomOutGeo, resetGeo, hoverClosestGeo

 1fig.show(config={"displayModeBar": True,
 2                 "displaylogo": False,
 3                 'modeBarButtonsToRemove': [
 4                     'zoom2d',
 5                     'toggleSpikelines',
 6                     'pan2d',
 7                     'select2d',
 8                     'lasso2d',
 9                     'autoScale2d',
10                     'hoverClosestCartesian',
11                     'hoverCompareCartesian']})

Line chart with "toggleSpikelines" "hoverCompareCartesian" buttons removed

Display a Plotly chart as a Static chart

A Plotly chart can be displayed as a static chart with the use of staticPlot parameter, which is false by default. Displaying the chart as a static chart automatically hides the Modebar as the plot is not interactive.

1fig.show(config={'staticPlot': True})

Plotly line chart displayed as static chart

Conclusion

Plotly is great for creating interactive charts that help bring the data to life. These interactive charts automatically default to displaying a toolbar on the top right called a "Modebar". This default Modebar can appear a bit cluttered and be confusing to viewers of the charts. The Modebar can be configured to always display or never display as well as limiting the buttons that are present in the Modebar.

Plotly - Display an Interactive Chart

Eric Callanan — Sunday, 20 Sep 2020

Pandas is great for searching, filtering and manipulating data and information can be visualised in charts easily using Matplotlib module. These charts are static and provide a snapshot of data that can be annotated to highlight certain data points. These are great for posters, slides and printed materials. It would be great if these charts could be more interactive as the world has moved to digital media and more data visualisations are viewed on screens. Plotly is used to recreate the under five mortality rate changes over time in an interactive chart.

Python modules used in this article

 1# import modules
 2import pandas as pd
 3import numpy as np
 4import shutil as sh
 5import requests
 6import os
 7
 8import plotly.express as px
 9import plotly.graph_objects as go
10from plotly.subplots import make_subplots

Plotly

Plotly is an open-source graphing library for Python that produces interactive charts. This allows the user to hover over data points to display exact values; Zoom in on sections of a chart; deselect key data series to focus on other data and reset the chart to the original. Plotly can be used to create extremely rich visualisations with many user interactions.

Load the data into a dataframe

How to download and load the excel data is covered in "Pandas - Load data from Excel file and Display Chart". This article will focus on creating charts with Plotly from the Pandas dataframe. The following code loads the data from Excel and filters to the median values. The source file is available from Unicef datasets at under-five mortality rate.

 1# Load the excel worksheet into a dataframe
 2u5mr_df = pd.read_excel(
 3    "/tmp/data/Under-five-mortality-rate_2020.xlsx",
 4    sheet_name = 'Country estimates (both sexes)',
 5    header = 14)
 6
 7# Drop the last two rows
 8u5mr_df.drop(u5mr_df.tail(2).index, inplace = True)
 9
10# Rename the columns to Years
11u5mr_df.columns = [x[:-2] if x.endswith('.5') else x for x in u5mr_df.columns]
12
13# Rename 'Uncertainty.Bounds*' column to 'Uncertainty.Bounds'
14u5mr_df = u5mr_df.rename(columns={'Uncertainty.Bounds*': 'Uncertainty.Bounds'})
15
16# Filter to the Median values
17u5mr_med_df = u5mr_df[u5mr_df['Uncertainty.Bounds'] == 'Median']
18
19# Review the data
20u5mr_med_df.iloc[[0,1,2,3,-4,-3,-2,-1], [0,1,2,3,4,5,6,-4,-3,-2,-1]]
21"""
22    ISO.Code Country.Name Uncertainty.Bounds  1950  1951  1952        1953       2016       2017       2018       2019
231        AFG  Afghanistan             Median   NaN   NaN   NaN         NaN  67.572190  64.940759  62.541196  60.269399
244        ALB      Albania             Median   NaN   NaN   NaN         NaN   9.419110   9.418052   9.525133   9.682407
257        DZA      Algeria             Median   NaN   NaN   NaN         NaN  24.792098  24.319482  23.805926  23.256168
2610       AND      Andorra             Median   NaN   NaN   NaN         NaN   3.369056   3.218925   3.085839   2.966929
27574      VNM     Viet Nam             Median   NaN   NaN   NaN         NaN  21.220796  20.843125  20.405423  19.935167
28577      YEM        Yemen             Median   NaN   NaN   NaN         NaN  56.823614  56.966430  58.460003  58.356138
29580      ZMB       Zambia             Median   NaN   NaN   NaN  234.418232  66.510929  64.337901  63.294182  61.663465
30583      ZWE     Zimbabwe             Median   NaN   NaN   NaN         NaN  59.538505  58.234924  55.856832  54.612967
31"""

Show simple Plotly chart

This code gets the top ten countries with the highest Under Five Mortality Rates in 2019 and uses plotly.graph_objects to create a simple barchart. While the layout of this chart could made nicer, hovering over each of the bars displays the details of the country and the U5MR for that country. The hover text displays external to the bar and automatically switches to displaying inside bar when there is not enough room. There is also a Plotly toolbar displayed on the top right of the chart. Selecting a region inside the chart will zoon in on the selected area.

 1Top_10 = u5mr_med_df.sort_values(by=u5mr_df.columns[-1], ascending=False)[["Country.Name", "2019"]].head(10)
 2"""
 3                 Country.Name        2019
 4376                   Nigeria  117.202078
 5481                   Somalia  116.972096
 6100                      Chad  113.790418
 797   Central African Republic  110.053912
 8466              Sierra Leone  109.236528
 9214                    Guinea   98.802973
10487               South Sudan   96.229299
11316                      Mali   94.035418
1255                      Benin   90.286429
1379               Burkina Faso   87.542426
14"""
15
16import plotly.graph_objects as go
17fig = go.Figure(
18    data = [go.Bar(x = list(Top_10["2019"]),
19                   y = list(Top_10["Country.Name"]),
20                   orientation = "h")],
21    layout = go.Layout(
22        title = go.layout.Title(
23            text = "Top ten countries with highest Under Five Infant Mortality in 2019"
24        )
25    )
26)
27
28fig.show()

Simple horizontal bar chart showing top ten countries with highest Under 5 Mortality Rates in 2019

Show top ten U5MR bar Chart

The following code modifies the default bar chart to improve the display. The chart layout, colors and font are changed as well as the format of the hover text.

 1Top_10 = u5mr_med_df.sort_values(by=u5mr_df.columns[-1], ascending=False)[["Country.Name", "2019"]].head(10)
 2bg_color = 'rgba(208,225,242,1.0)'
 3
 4fig = go.Figure(
 5    data = [go.Bar(x = list(Top_10["2019"]),
 6                   y = list(Top_10["Country.Name"]),
 7                   hovertemplate = "%{y}: <br><br>U5MR: %{x:.1f}<extra></extra>",
 8                   orientation = 'h')]
 9)
10
11fig.update_layout(
12    # Set default font
13        font = dict(
14        family = "Droid Sans",
15        size = 16
16    ),
17    # Set figure title
18    title = dict(
19        text = "<b>Top ten countries with highest Under Five Mortality Rate in 2019</b>",
20        xref = 'container',
21        yref = 'container',
22        x = 0.5,
23        y = 0.91,
24        xanchor = 'center',
25        yanchor = 'middle',
26        font = dict(family = 'Droid Sans', size = 24)
27    ),
28    # set x-axis
29    xaxis = dict(
30        title = dict(
31            text = 'Under Five Mortality Rate (per 1000 live births)',
32            font = dict(family = 'Droid Sans', size = 18)
33        ),
34        showgrid = False,
35        linecolor = bg_color,
36        linewidth = 2,
37        showticklabels = True,
38    ),
39    # set y-axis
40    yaxis = dict(
41        showgrid = False,
42        linecolor = bg_color,
43        linewidth = 4,
44    ),
45    # set the plot bacground color
46    plot_bgcolor = bg_color,
47    # set the hover background color
48    hoverlabel = dict(
49        bgcolor = 'rgba(75,152,201,0.2)',
50        font_size = 16
51    ),
52    paper_bgcolor = bg_color,
53)
54
55fig.update_traces(marker_color = 'rgb(75,152,201)')
56fig['layout']['yaxis']['autorange'] = "reversed"
57
58fig.show()

Nicer bar chart showing top ten countries with highest Under 5 Mortality Rates in 2019

Show top ten U5MR over time

There is a limit to how beneficial an interactive chart is when it is just displaying the total numbers for the top ten countries in the bar chart above. The interactive chart is much more useful when looking at a multiple sets of data plotted over time such as the changes in the current top ten countries over time. The following line chart shows the changes in Under 5 Mortality Rate for countries with the highest rates in an interactive plotly chart. This is similar to a static chart that was previously created with Matplotlib. The advantages are that exact numbers fo a prticular can be seen by hovering over that point as well as hiding some countries to focus on countries of interest.

This function Extract the mortality rates for the top number of countries based on rates in the latest year.

 1def get_top_countries(data_df, lowest = True, num = 10):
 2    '''Extract the mortality rates for the top number of countries
 3    based on rates in the latest year
 4
 5    Keyword arguments:
 6    data_df  -- dateframe of all the mortality rates for all the countries
 7    lowest   -- boolean flag to either contries with the lowest rate
 8                or highest rate (default is True)
 9    num      -- number of countries to return (default is 10)
10
11    return: dataframe that has been transposed and filtered to the top n counbtries
12    '''
13    # Need to transpose the dataframe to use year as the x-axis
14    df = data_df.sort_values(by = data_df.columns[-1], ascending = lowest).head(10).T
15    df.reset_index(drop = False, inplace = True)
16
17    # Set the Country.Name as the heading for the columns
18    df.columns = df.iloc[np.where(df['index'] == 'Country.Name')[0][0]]
19
20    # Rename the Country.Name column to Year
21    df = df.rename(columns = {'Country.Name': 'Year'})
22
23    # Drop the rows that do not contain u5mr data
24    df = (df[df['Year']
25             .isin(['ISO.Code',
26                    'Country.Name',
27                    'Uncertainty.Bounds']) == False]
28        )
29    df.reset_index(drop = True, inplace = True)
30
31    return df

This function creates the Plotly line chart showing the changes in each country over time. It uses a separate function to format the hover text to make it easier to modify and maintain.

  1def hov_text(ary_x, ary_y, country):
  2    '''Create the hover text for each data point
  3
  4    Keyword arguments:
  5    ary_x    -- list of all the x data points
  6    ary_y    -- list of all the y data points
  7    country  -- name of country
  8
  9    return: A list of hover texts for the selected country
 10            with format to display
 11    '''
 12    txt = [f'''
 13    <b>{country}<b><br><br>
 14    Year: <b>{ary_x[i]}</b><br>
 15    U5MR: <b>{ary_y[i]:.1F}</b><br>
 16    <extra></extra>
 17    ''' for i in range(len(ary_x))]
 18    return txt
 19
 20
 21def create_top_case_chart(df, title):
 22    '''Create a plotly line chart from the dataframe, which must contain
 23    a column for 'Year' and the other columns as countries
 24
 25    Keyword arguments:
 26    df      -- dateframe of the mortality rates for all the countries
 27    title   -- text to be displayed as the title of the chart
 28
 29    return: a Plotly Graph Object figure
 30    '''
 31    bg_color = 'rgba(208, 225, 242, 1.0)'
 32    line_color = 'rgba(75, 152, 201, 1.0)'
 33    grid_color = 'rgba(75, 152, 201, 0.3)'
 34
 35    y_max = ((max(df.drop(['Year'], axis = 'columns').max()) // 100) + 1) * 100
 36    y_min = 0
 37    x_min = int(min(df['Year'])) - 2
 38    x_max = int(max(df['Year'])) + 2
 39
 40    fig = go.Figure()
 41    for c in df.drop(['Year'], axis = 'columns').columns:
 42        fig.add_trace(go.Scatter(x = df.Year,
 43                                 y = df[c],
 44                                 name = c,
 45                                 hovertemplate = hov_text(df.Year, df[c], c)))
 46
 47    fig.update_traces(
 48        hoverinfo = 'text+name',
 49        mode = 'lines'
 50    )
 51
 52    fig.update_layout(
 53        # Set figure title
 54        title = dict(
 55            text = f'<b>{title}</b>',
 56            xref = 'container',
 57            yref = 'container',
 58            x = 0.5,
 59            y = 0.9,
 60            xanchor = 'center',
 61            yanchor = 'middle',
 62            font = dict(family = 'Droid Sans', size = 28)
 63        ),
 64        # set legend
 65        legend = dict(
 66            orientation = "h",
 67            traceorder = 'normal',
 68            font_size = 12,
 69            x = 0.0,
 70            y = -0.3,
 71            xanchor = "left",
 72            yanchor = "top"
 73        ),
 74        # set x-axis
 75        xaxis = dict(
 76            title = 'Year',
 77            range = [x_min, x_max],
 78            linecolor = line_color,
 79            linewidth = 2,
 80            gridcolor = grid_color,
 81            showticklabels = True,
 82            ticks = 'outside',
 83        ),
 84        # set y-axis
 85        yaxis = dict(
 86            title = 'Under Five Mortality Rate',
 87            range = [y_min, y_max],
 88            linecolor = line_color,
 89            linewidth = 2,
 90            gridcolor = grid_color,
 91            showticklabels = True,
 92            ticks = 'outside',
 93        ),
 94        showlegend = True,
 95        # set the plot bacground color
 96        plot_bgcolor = bg_color,
 97        paper_bgcolor = bg_color,
 98    )
 99
100    return fig

Finally these functions are used to create the chart for the top ten countries.

1top_10_fig = create_top_case_chart(
2    get_top_countries(u5mr_med_df, False),
3    f'Changes in Under Five Mortality Rate <BR>for countries with the highest in {u5mr_df.columns[-1]}')
4top_10_fig.show()

Interactive chart showing changes in Under 5 Mortality Rate for countries with the highest rates

Show lower ten U5MR over time

It can take some time formatting the layout and colors of a chart to display just right. The good news is that the same functions can be used to display similar sets of data. The following code is all that is needed to get the ten countries with the current lowest Under Five Mortality Rates.

1lowest_10_fig = create_top_case_chart(
2    get_top_countries(u5mr_med_df, True),
3    f'Changes in Under Five Mortality Rate <BR>for countries with the lowest in {u5mr_df.columns[-1]}')
4lowest_10_fig.show()

Interactive chart showing changes in Under 5 Mortality Rate for countries with the lowest rates

Display data for specific countries

The same chart function can be used to display the changes on mortality rates over time. The function above to retrieve the top five is a little too specific and could possibly split into two functions - one to filter the data of interest and a second to transpose the data and prepare it for creating the chart.

 1countries = [
 2    'Canada', 'China', 'France', 'Iceland',
 3    'India', 'Ireland', 'Mexico', 'Sweden',
 4    'United Kingdom', 'United States of America'
 5]
 6
 7select_df = u5mr_med_df[u5mr_med_df['Country.Name'].isin(countries)]
 8
 9# Need to transpose the dataframe to use year as the x-axis
10select_df = select_df.sort_values(by = select_df.columns[-1]).T
11select_df.reset_index(drop = False, inplace = True)
12
13# Set the Country.Name as the heading for the columns
14select_df.columns = select_df.iloc[np.where(select_df['index'] == 'Country.Name')[0][0]]
15
16# Rename the Country.Name column to Year
17select_df = select_df.rename(columns = {'Country.Name': 'Year'})
18
19# Drop the rows that do not contain u5mr data
20select_df = (select_df[select_df['Year']
21                       .isin(['ISO.Code',
22                              'Country.Name',
23                              'Uncertainty.Bounds']) == False]
24            )
25select_df.reset_index(drop = True, inplace = True)
26
27select_fig = create_top_case_chart(
28    select_df,
29    f'Changes in Under Five Mortality Rate for selected countries')
30
31select_fig.show()

Interactive chart showing changes in Under 5 Mortality Rate for selected countries

Host the interactive chart in static web site

I've created dashboards using Dash, which creates nice interactive dashboards in flask-like apps. I'd like to give a shout out to Igor Gotlibovych for providing instructions on how to host plotly interactive charts on a Hugo static website - Including plotly figures in Hugo posts. Thank you for providing this information.

Conclusion

Matplotlib is great for creating charts in Python, but the results are static images of the data. Plotly is great for creating interactive charts that help bring the data to life. These charts can be customised in a number of ways to present data that users can hover over to see more information, hide some data by deselecting on the legend or zoom in on a section of the chart.

plotly on Software Development Notes

Countries with highest deaths from covid

Load Covid data

Load Vaccination data

Load population data

Compare countries in the datasets

Merge datasets on country

Overview of Countries with highest deaths from covid per capita

Vaccination status in countries with highest deaths

Conclusion

Daily counts and rolling average of COVID-19 numbers

Load COVID-19 data into dataframes

Get the top 10 countries

Display the deaths for a country over time

Display daily counts for a single country

Display rolling average on daily counts

Identify the best rolling average window size

Display rolling averages for top ten countries

Display rolling averages and the daily counts

Conclusion

Display COVID-19 numbers relative to population

Load COVID-19 data into dataframes

Load global population data into dataframe

Load population data for a single year

Compare countries in COVID-19 datasets with population dataset

Update country names in population dataset

Calculate cases and deaths per 100,000 of population

Countries with highest confirmed cases per 100,000 of population

Countries with highest deaths per 100,000 of population

Conclusion

Add data with a secondary y axis

Load the data into a Pandas dataframe

Global confirmed cases of COVID-19

Global deaths from COVID-19

Show Confirmed cases and deaths from COVID-19 on a single graph

Display data for selected country

Conclusion

Countries with highest confirmed cases of Covid-19 with Plotly

Load the data into a Pandas dataframe

Display confirmed cases in top ten countries on the latest date

Display confirmed cases in US

Create a function to display confirmed cases for any country

Display confirmed cases for top ten countries

Conclusion

Plotly - Customise plotly toolbar

Show or hide the Modebar

Create the sample line chart

Always show the Modebar

Never show the Modebar

Remove the plotly logo from the Modebar

Remove specific buttons from the Modebar

Chart

Button names

Display a Plotly chart as a Static chart

Conclusion

Plotly - Display an Interactive Chart

Python modules used in this article

Plotly

Load the data into a dataframe

Show simple Plotly chart

Show top ten U5MR bar Chart

Show top ten U5MR over time

Show lower ten U5MR over time

Display data for specific countries

Host the interactive chart in static web site

Conclusion