clean_daily.py

import math
# unique dates 
dates = vaccine_df.date.unique().tolist()
#For more smoother animations
dates.extend(['2020-12-12', '2020-12-13']) 

# unique countries 
countries = vaccine_df.country.unique().tolist()

# for easy processing 
short = vaccine_df[['date', 'country', 'total_vaccinations']]

# values of unqiue (date, country) already in short 
# i.e we want to make sure we have some data for each, even if it is 0 
keys= list(zip(short.date.tolist(), short.country.tolist()))
for date in dates:
    for country in countries:
        idx = (date, country)
        if idx not in keys:
            if date == min(dates):
                # this means there's no entry for {country} on the earliest date 
                short = short.append({
                    "date": date, 
                    "country": country, 
                    "total_vaccinations": 0
                }, ignore_index=True)
            else:
                # entry for {country} is missing on a date other than the earliest
                short = short.append({
                    "date": date, 
                    "country": country, 
                    "total_vaccinations": pd.NA
                }, ignore_index=True)
                
#fill missing values with previous day values (this is OK since it is cumulative)
short = short.sort_values(['country', 'date'])

short.total_vaccinations = short.total_vaccinations.fillna(method='ffill')

# scale the number by log to make the color transitions smoother
vaccines = short.sort_values('date')
vaccines['log_scale'] = vaccines['total_vaccinations'].apply(lambda x : math.log2(x+1))