Utility for comparing COVID-19 cases side by side between countries.

README.md

A small CLI tool for comparing COVID-19 numbers between countries. The data is fetched from https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data and cached for 12 hours.

Use convid.py -h for more information on how to use the tool.

covid.py

#!/usr/bin/python3
#
# Source: https://gist.github.com/bysse/4959eae4f99a934dfc75a49eb7667d2b
#
import argparse
import os.path
import sys
import time
import urllib.request

try:
    plot_available = True
    import matplotlib.pyplot as plt
except:
    print('WARNING: Missing matplotlib, no plot available')
    plot_available = False

URL_BASE = 'https://github.com/CSSEGISandData/COVID-19/raw/master/csse_covid_19_data/csse_covid_19_time_series/'
CONFIRMED = 'time_series_covid19_confirmed_global.csv'
DEATHS = 'time_series_covid19_deaths_global.csv'
RECOVERED = 'time_series_covid19_recovered_global.csv'

SEPARATOR = ' | '


def filedate(filename):
    if not os.path.exists(filename):
        return 0
    stat = os.stat(filename)
    try:
        return stat.st_birthtime
    except AttributeError:
        return stat.st_mtime


def verify_data_file(filename):
    if time.time() - filedate(filename) < 12 * 60 * 60:
        return
    if os.path.exists(filename):
        os.remove(filename)
    print("# Downloading", filename)
    urllib.request.urlretrieve(URL_BASE + filename, filename)


parser = argparse.ArgumentParser(description='Compare covid-19 data between countries')
parser.add_argument('countries', metavar='country', nargs='+',
                    help='countries to display data for')
parser.add_argument('-t', '--threshold', metavar='N', dest='threshold', type=int, default=1,
                    help='the minimum number of cases to start comparison from')

if plot_available:
    parser.add_argument('-p', '--plot', action='store_true', help='Create and show a plot')

group = parser.add_mutually_exclusive_group(required = False)
group.add_argument('--sick', action='store_true', help='Use the number of sick as the data set.')
group.add_argument('--dead', action='store_true', help='Use the number of dead as the data set.')
group.add_argument('--recovered', action='store_true', help='Use the number of recovered as the data set.')

args = parser.parse_args()

data_set = CONFIRMED
plot_title = "Confirmed Cases"

if args.dead:
    data_set = DEATHS
    plot_title = "Deaths"
if args.recovered:
    data_set = RECOVERED
    plot_title = "Recoveries"

verify_data_file(data_set)

countries = []
for country in args.countries:
    countries.append(country.lower())
countries.sort()

data = {}
start = {}
legend = None
with open(data_set, 'r') as fp:
    for line in fp.readlines():
        fields = line.split(',')

        if not legend:
            legend = list(map(lambda x: x.strip(), fields[4:]))
            legend = list(map(lambda x: x[:-3], legend))
            continue

        country = fields[1].lower()
        if country in countries:
            cases = list(map(int, fields[4:]))
            data[country] = cases

            # find the first index
            start[country] = len(cases)
            for n in range(len(cases)):
                if cases[n] >= args.threshold:
                    start[country] = n
                    break

# calculate column width
columns = [1 + max(map(len, legend))]
for country in countries:
    columns.append(len(country))

column_format = []
for col in columns:
    column_format.append( "%{}s".format(col) )


min_index = min(start.values())

def output(index, value, more_fields=True):
    print(column_format[index] % value, end='')

    if more_fields:
        print(SEPARATOR, end='')
    else:
        print()

plot_data = {}

num_columns = len(countries)
for i in range(num_columns):
    output(0, 'date')
    output(1 + i, countries[i], i+1 != num_columns)
    plot_data[countries[i]] = []

for index in range(min_index, len(legend)):
    i = 0
    for country in countries:        
        offset = start[country] - min_index + index
        new_line = i+1 != num_columns

        if offset >= len(legend):
            output(0, '')
            output(1+i, '', new_line)
            plot_data[country].append(None)
        else:
            output(0, legend[offset])
            output(1 + i, data[country][offset], new_line)
            plot_data[country].append(data[country][offset])

        i += 1

if 'plot' in args and args.plot:
    for country, data in plot_data.items():
        plt.plot(data, label=country)
    
    plt.title("COVID-19")
    plt.ylabel(plot_title)
    plt.xlabel("Days")
    plt.legend(loc="upper left")
    plt.grid()
    plt.show()