Github Classroom Analytics

A python tool for tracking student engagement in github classroom assignments.

Essentially it uses the classroom roster and the github API to check whether students have set up a repository and if they have, it grabs their commit data and generates basic statistics.

Basic usage

A simple example which loops over all the students and prints out their stats.

from github_classroom.config import assignment_from_config

assignment = assignment_from_config('config.ini')

for identifier, github_username, repo in assignment.roster():
    if repo:
        print()
        print(f'{identifier!r}: {github_username}')
        commits = repo.commit_count_by_week()
        for week in sorted(list(commits)):
            print(f'\t{commits[week]} commits in the week beginning {week}')
        print()
    else:
        print(f'{identifier!r} has not created a repo yet')

The above code generates output in the following format:

'student1' has not created a repo yet
'student2' has not created a repo yet
'student3' has not created a repo yet

'student4': 'myAmazingGithubName'
	1 commits in the week beginning 2021-09-26
	4 commits in the week beginning 2021-10-03
	6 commits in the week beginning 2021-10-10
	2 commits in the week beginning 2021-10-17
	35 commits in the week beginning 2021-10-24

See the csv_example for a more practical approach

Configuration

The above example requires a 'config.ini' file like this:

[github]
username = myGithubUsername
token = ghp_personal_access_token
organisation = GithubOrganisationName

[classroom]
assignment_prefix = myassignment
roster_filename = classroom_roster.csv

Note that the roster file itself will need to be updated regularly to bring in new github_usernames as students accept the assignment.

Installation

Clone the git repository, set up a virtual environment as required and run this to install the github_classroom library.

pip install --editable .

The --editable flag allows for further updates to be pulled in by git. This will be necessary whilst we are in the very early stages of development.

Note the full stop in the above command, the command should be run in the directory where the code was cloned.