New confirmed COVID-19 cases per county over time

new_confrmed_cases.py

from datetime import datetime

import pandas as pd
import requests

def get_usafacts_county_cases(url):
    r = requests.get(url)
    return pd.read_csv(io.BytesIO(r.content))
    
def get_new_cases(url):
    def isdate(x):
        try:
            datetime.strptime(x, '%m/%d/%Y')
            return True
        except:
            return False
    
    df = get_usafacts_county_cases(url)
    date_columns = list(filter(lambda x: isdate(x), df.columns))
    total_cases = df.set_index('countyFIPS')[date_columns].copy()
    x = pd.concat([total_cases, -1 * total_cases.shift(1, axis=1)])
    new_cases = x.reset_index().groupby('countyFIPS').sum().reset_index()

    df = df.drop(columns=date_columns).merge(new_cases, on='countyFIPS')
    df = df[df['countyFIPS'] != 0]
    return df

new_confirmed_cases_per_county = get_new_cases(
    'https://static.usafacts.org/public/data/covid-19/covid_confirmed_usafacts.csv',
)

new_confirmed_cases_per_county