Plot word frequency over time in a git repo's commit messages using matplotlib.

word_frequency.py

#!/usr/bin/python

"""
Plot the frequency of the given words in the commit messages of a git repository.
  Usage:    python word_frequency.py git_dir word1 [word2...]
  Example:  python word_frequency.py /path/to/repo/.git fix add remove
"""

import sys
import datetime
import pygit2
import matplotlib.pyplot as plot

def month_for_date(date):
    return datetime.datetime.strptime(date.strftime("%Y %m"), "%Y %m")

if len(sys.argv) < 3:
    print __doc__
    sys.exit(1)

repo = pygit2.Repository(sys.argv[1])
keywords = set(sys.argv[2:])

dates_freqs = []
for commit in repo.walk(repo.head.oid, pygit2.GIT_SORT_TIME):
    date = datetime.date.fromtimestamp(commit.commit_time)
    freq = dict(zip(keywords, map(commit.message.lower().count, keywords)))
    dates_freqs.append((date, freq))

for keyword in keywords:
    data = {}
    for date, freq in dates_freqs:
        month = month_for_date(date)
        if not data.get(month):
            data[month] = 0
        data[month] += freq[keyword]

    keys, vals = [], []
    for key in sorted(data.keys()):
        keys.append(key)
        vals.append(data[key])

    plot.plot(keys, vals, label=keyword)

plot.legend()
plot.show()