nba_travel_impact_get_data.py

# Get all team totals for all dates
temp_df_list = list()
for game_date in game_dates:
    logger.info(f'Getting data for games on {game_date}')
    temp_boxes = client.team_box_scores(day=game_date.day, month=game_date.month, year=game_date.year)
    temp_df = pd.DataFrame(temp_boxes)
    temp_df = temp_df.assign(game_date=str(game_date))
    temp_df_list.append(temp_df)

# ========== SET UP BOX SCORE DATAFRAME ==========

box_df = pd.concat(temp_df_list)
box_df.reset_index(drop=True, inplace=True)

box_df = box_df.assign(home=True)
for i, df_row in box_df.iterrows():
    gm_date = df_row.game_date
    tm_str = df_row.astype(str)['team'][5:]
    sch_row = schedule_df[
        ((schedule_df.home_team == tm_str) | (schedule_df.away_team == tm_str))
        & (schedule_df.game_date == pd.to_datetime(gm_date).date())
    ]
    if sch_row.iloc[0].away_team == tm_str:
        logger.info(f'Setting {tm_str} game on {gm_date} as an away game.')
        box_df.loc[i, 'home'] = False

box_df.to_csv('srcdata/' + str(season_end_yr) + '_nba_team_box_raw.csv')

# ========== PROCESS DATA BASED ON TRAVEL & REST DAYS ==========

box_df = box_df.assign(rest_days=10)
box_df = box_df.assign(travelled=False)

for temp_tm in box_df.team.unique():
    # temp_tm = Team.TORONTO_RAPTORS
    team_box_df = box_df[(box_df.team == temp_tm)]

    # team_box_df.reset_index(drop=True, inplace=True)
    for i in range(len(team_box_df)):
        tmp_ind = team_box_df.iloc[i].name
        if i > 0:
            df_row = team_box_df.iloc[i]
            temp_td = pd.to_datetime(team_box_df.iloc[i]['game_date']) - pd.to_datetime(team_box_df.iloc[i-1]['game_date'])
            box_df.loc[tmp_ind, 'rest_days'] = temp_td.days - 1
            if df_row.home == False:
                box_df.loc[tmp_ind, 'travelled'] = True
            elif team_box_df.iloc[i-1].home == False:
                box_df.loc[tmp_ind, 'travelled'] = True

box_df = box_df.assign(rest_cats=box_df.rest_days.clip(upper=2).astype(str).replace('2', '2+'))
box_df = box_df.assign(win=box_df.outcome == Outcome.WIN)

box_df.team = box_df.team.astype(str).str[5:]
box_df.outcome = box_df.outcome.astype(str).str[8:]

box_df.to_csv('srcdata/' + str(season_end_yr) + '_nba_team_box_scores.csv')