URI Matching

spotify-wrapped.py

track_artist_album_df['Artist Image URL'] = track_artist_album_df['Artist URI']
    .map(unique_artist_image_url_dict)
spotify_data = pd.merge(cleaned_df, track_artist_album_df, on='Track URI', how='left')

def get_top_100(entity, spotify_data):
    
    if entity == 'Track':
        spotify_data[['Track', 'Album']] = spotify_data[['Track', 'Album']]
            .apply(lambda x: x.str.replace('Feat', 'feat'))
        track_artist_groups = spotify_data.groupby(['Track', 'Artist'])
        songs_dict = {}
        for group, group_df in track_artist_groups:
            max_uri = group_df['Track URI'].value_counts().idxmax()
            group_df['Track URI'] = max_uri
            songs_dict[group] = group_df
        spotify_data = pd.concat(songs_dict.values(), ignore_index=True)
        
    uri_counts = spotify_data[f'{entity} URI'].value_counts().reset_index(name='Streams')
    uri_counts.columns = [f'{entity} URI', 'Streams']
    spotify_data = pd.merge(spotify_data, uri_counts, on=f'{entity} URI', how='left')
    spotify_data = spotify_data.drop_duplicates(subset=f'{entity} URI')
    spotify_data.sort_values(by='Streams', ascending=False, inplace=True, ignore_index=True)
    top_100 = spotify_data.head(100).copy().assign(Rank=lambda x: range(1, len(x) + 1))
    top_100['Streams'] = top_100['Streams'].astype(int)
    return top_100