Counts of users' events by day

active_user_analysis.py

"""
users if active or not on a given day plus counts of events
"""

import os
import csv
from collections import defaultdict
from collections import Counter

UFILE = 'data/users.csv'
EFILE = 'data/events.csv'
OUTFILE = 'out/third.csv'

def csv_to_list(fn):
    with open(fn) as f:
        reader = csv.reader(f)
        return list(reader)


if __name__ == '__main__':
    events = csv_to_list(EFILE)
    users = csv_to_list(UFILE)

    all_dates = {e[2].split()[0] for e in events[1:]}
    all_uids = {u[0] for u in users[1:]}
    all_events = {e[1] for e in events[1:]}

    user_dates = defaultdict(lambda: defaultdict(Counter))

    for e in events[1:]:
        user_dates[e[0]][e[2].split()[0]][e[1]] += 1

    # user_dates userid -> date -> event -> count
    sorted_etypes = sorted(e for e in all_events)
    count_headers = [f'count_{e}' for e in sorted_etypes]
    event_string = ','.join(count_headers)

    outlines = [f'user_id,date,active,{event_string}\n']

    for uid in sorted(all_uids):
        for date in sorted(all_dates):
            try:
                active = 't' if date in user_dates[uid] else 'f'
            except KeyError:
                active = 'f'

            count_string = ','.join(str(user_dates[uid][date][e]) for e in
                                    sorted_etypes)
            outlines.append(','.join([uid, date, active, count_string]) +'\n')

    os.makedirs(os.path.dirname(OUTFILE), exist_ok=True)
    with open(OUTFILE, 'w') as f:
        f.writelines(outlines)