cal/paper-dynasty-card-creation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added PA and AB to batter_stats

Browse Source
This commit is contained in:
Cal Corum 2024-10-17 16:31:17 -05:00
parent 07faea0bc7
commit 1109a12434
2 changed files with 607 additions and 588 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1172
batting_stats.csv
View File

File diff suppressed because it is too large Load Diff

23
retrosheet_data.py
View File

@ -12,11 +12,11 @@ date = f'{datetime.datetime.now().year}-{datetime.datetime.now().month}-{datetim
log_level = logging.INFO
logging.basicConfig(
filename=f'logs/{date}.log',
format='%(asctime)s - card-creation - %(levelname)s - %(message)s',
format='%(asctime)s - retrosheet_data - %(levelname)s - %(message)s',
level=log_level
)
FILE_PATH = 'data-input/retrosheet/'
EVENTS_FILENAME = 'retrosheets_events_1998_short.csv'
EVENTS_FILENAME = 'retrosheets_events_1998_short.csv' # Removed last few columns which were throwing dtype errors
PERSONNEL_FILENAME = 'retrosheets_personnel.csv'
@ -39,6 +39,17 @@ def get_batting_stats_by_date(start_date: int, end_date: int) -> pd.DataFrame:
bs['batter_id'] = all_plays['batter_id'].unique()
bs = bs.set_index('batter_id')
pal_series = all_plays[(all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'l')].groupby('batter_id').count()['event_type'].astype(int).rename('PAvL')
bs = pd.concat([bs, pal_series], axis=1)
par_series = all_plays[(all_plays.batter_event == 't') & (all_plays.pitcher_hand == 'r')].groupby('batter_id').count()['event_type'].astype(int).rename('PAvR')
bs = pd.concat([bs, par_series], axis=1)
abl_series = all_plays[(all_plays.ab == 't') & (all_plays.pitcher_hand == 'l')].groupby('batter_id').count()['event_type'].astype(int).rename('ABvL')
bs = pd.concat([bs, abl_series], axis=1)
abr_series = all_plays[(all_plays.ab == 't') & (all_plays.pitcher_hand == 'r')].groupby('batter_id').count()['event_type'].astype(int).rename('ABvR')
bs = pd.concat([bs, abr_series], axis=1)
# Basic counting stats
for event_type, vs_hand, col_name in [
('home run', 'r', 'HRvR'),
@ -75,8 +86,16 @@ def get_batting_stats_by_date(start_date: int, end_date: int) -> pd.DataFrame:
async def main(args):
print(f'Running the calcs...')
start = datetime.datetime.now()
data = get_batting_stats_by_date(start_date=101, end_date=430)
end_calc = datetime.datetime.now()
print(f'Saving to csv...')
data.to_csv(f'batting_stats.csv')
end = datetime.datetime.now()
print(f'Done!\n\nCalc time: {(end_calc - start).total_seconds()}s\nSave time: {(end - end_calc).total_seconds()}s\nTotal: {(end - start).total_seconds()}s')
if __name__ == '__main__':