starbucks_post_6.py

transcript_portfolio = pd.merge(transcript, portfolio, left_on='offer_id', right_on='id', how='left')
transcript_by_group = transcript_portfolio.groupby(['person', 'offer_id'])

completion_details = [];

'''
Go through each group in the transaction grouping. Because iterating can be slow, 
we will use vectorized operations inside the main loop.
'''
for i, g in transcript_by_group:
    
    record = {};
    
    # Record status based on weather the offer has been viewed
    if g[['event_offer_received', 'event_offer_viewed', 'event_offer_completed']].sum(axis=0).sum() == 3:
        record['status'] = 'viewed_and_completed';
    elif (g[['event_offer_received', 'event_offer_completed']].sum(axis=0).sum() == 2) and 'status' not in record:
        record['status'] = 'not_viewed_and_completed';
    else:
        record['status'] = 'no_offer';
    
    # Get required details
    person_id, offer_id = g['person'].iloc[0], g['offer_id'].iloc[0];
    first_purchase, last_purchase = g['time'].min(), g['time'].max();
    
    # Get all transactions corresponding to the person
    try:
        person_transactions = transactions_only.get_group(person_id);
    except KeyError:
        offer_transactions = [];
    
    # Filter out transactions that are within the time window of the offer
    offer_transactions = person_transactions[(person_transactions.time > first_purchase) \
                                             & (person_transactions.time < last_purchase)];

    # Store these details into the dictionary
    record['num_transactions'] = len(offer_transactions);
    record['person_id'], record['offer_id'] = person_id, offer_id;
    record['first_purchase'], record['last_purchase'] = first_purchase, last_purchase;
    
    #print(person_id, offer_id, num_transactions)
    completion_details.append(record);

completion_details_df = pd.DataFrame(completion_details);