1999-2002 Run Expectancy (from CSV)

run_exp.py

import csv
import glob
import os
import sys

from mattreduce import Job


get_state = lambda outs, bases: '%s:%s' % (outs, bases)


def reader():
    path = os.path.join(sys.argv[1], '*.csv')

    for fn in glob.glob(path):
        reader = csv.DictReader(open(fn, 'r'))

        for row in reader:
            # skip partial inning events
            if (int(row['INN_CT']) >= 9
                and row['HOME_TEAM_ID'] == row['BAT_TEAM_ID']):
                continue

            yield row


def mapper(row):
    key = get_state(row['OUTS_CT'], row['START_BASES_CD'])
    value = int(row['EVENT_RUNS_CT']) + int(row['FATE_RUNS_CT'])

    yield key, value


def reducer(key, values):
    return key, (sum(values) / float(len(values)))


def main():
    for k, v in Job(mapper, reducer).run(reader()):
        print k, round(v, 3)


if __name__ == '__main__':
    main()