GitHub statistics

This project downloads data about a given public GitHub repository and helps to visualize some statistics with different plots. The data is downloaded from the official GitHub API

Usage

1. Install the dependencies with pip install -r requirements.txt
2. (Optional) Set the GITHUB_OAUTH_TOKEN environment variable to your personal GitHub Oauth token (GitHub - Creating a personal access token). You can use the dotenv package and a .env file like in the examples.
3. Instantiate a new Visualizer and specify the owner and repository parameters.

pandas_visualizer = Visualizer('porsk', 'github-stats')

4. Use the provided methods for plotting statistics about the repository.

Note: If you have not turned the cache functionality off with the useCache parameter of the visualizer subsequent calls should be faster.

Examples

You can find examples for the usage in the Downloader Example and Visualizer Example notebooks.

Downloader (hub_downloader.py)

• expects an owner and a repository name
• optionally a githab oauth token (by default as an anonymous user the limit is 100 request per hour, but with a token 5000)
• it caches the downloadad data by default into the data directory
• can be forsed to ignore the cache and redownload the data with the useCacheIfAvailable parameter
• by default it logs some information about the status, but it can be disabled with the verbose paramter

Methods for fetching data

get_contributors_statistic()

Returns two lists, one with the total commit counts by user and another list with additions, deletions, and commit counts by week and user.

Total contributions example:

index	user	commits
0	xhochy	10
1	Bharat123rox	10
...	...	...
98	jreback	2758
99	wesm	2994

Weekly contributions example:

index	user	week_unix_ts	date	additions	deletions	commits
0	xhochy	1249171200	2009-08-02	0	0	0
1	xhochy	1249776000	2009-08-09	0	0	0
...	...	...	...	...	...	...
56398	wesm	1589068800	2020-05-10	0	0	0
56399	wesm	1589673600	2020-05-17	0	0	0

get_code_frequency_statistic()

Returns a weekly aggregate of the number of additions and deletions pushed to the repository.

Example:

index	week_unix_ts	additions	deletions	date
0	1249171200	21659	-4	2009-08-02
1	1249776000	0	0	2009-08-09
...	...	...	...	...
562	1589068800	7253	-5247	2020-05-10
563	1589673600	1723	-986	2020-05-17

get_issues()

Returns the list of open issues in the repository.

Example:

index	id	state	created_at
0	621098694	open	2020-05-19
1	621095664	open	2020-05-19
...	...	...	...
3542	4485088	open	2012-05-09
3543	4217456	open	2012-04-20

get_commit_activity()

Returns the last year of commit activity grouped by week.

Example:

index	week_unix_ts	mon	tue	wed	thu	fri	sat	sun	week
0	1558828800	2	4	5	6	1	15	4	2019-05-26
1	1559433600	7	2	11	5	10	9	4	2019-06-02
...	...	...	...	...	...	...	...	...	...
50	1589068800	15	14	12	2	7	1	14	2020-05-10
51	1589673600	10	7	0	0	0	0	11	2020-05-17

get_stargazers()

Returns the lists of people that have starred the repository.

Example:

index	user	starred_at
0	sbusso	2010-08-24
1	auser	2010-08-24
...	...	...
24975	Badboy-16	2020-05-19
24976	ejungwoo	2020-05-19

Visualizer (hub_visualizer.py)

• expects an owner and a repository name
• uses the downloader in the background for getting the data
• by default it will try to get the GitHub oauth token from the GITHUB_OAUTH_TOKEN environment variable

Methods for plotting graphs

commit_activity()

Plots a grid plot about the commit activity in the repository during the last year.

Commit activity example:

lines_over_time()

Plots two graphs, one showing the total lines of code over time, the other the additions and deletions over time using line charts.

Total lines over time example:

Additions and deletions over time example:

commits_by_author()

Plots a pie chart showing the top contributors based on the commit count. With the optional limit parameter the number of shown contributor can be modified.

Commits by author example:

stargazer_history()

Plots two line charts, one showing the number of stars on the repo over time and the other showing the number of new stars month by month.

Number of stars over time example:

New stars aggregated by months example: