Accessing and storing GitHub traffic data automatically

Does anyone have any recommendations for tools/metrics for tracking package use/success? For starters, I’d like to track GitHub traffic, since only the last 14 days are available within GitHub. I know this is possible with the GitHub API, and I’ve found multiple tools to do so, but I’m wondering if anyone has experience with/would recommend any that are currently being maintained?

Beyond those statistics, we’ve got workshop survey data, but are looking for other suggestions on ways to best track our reach (since obviously GitHub statistics cannot tell the full story if people are continuing to use the software or access it through a shared environment (i.e. don’t download it directly).

2 Likes

You may find this blog post by Matt Rocklin useful: Estimating Users

Thanks @andersy005! That was a great read, and I’m probably going to use some of the pypistats functionality.

Any recommendations for a library/tool/GitHub action to track GitHub statistics?

Thanks @andersy005! That was a great read, and I’m probably going to use some of the pypistats functionality.

You’re welcome!

If you’re interested in

  • the “git” part of the project, git-quick-stats is a great command line tool. Of course, this won’t tell you much about the traffic (such as the repo visits, etc) on GitHub.

  • the “GitHub” part (i.e. stats on community activity), give repository-traffic-action a try. I haven’t used it myself though :slight_smile:

Also, when you get a chance, check this project: Github Statistics out and see if it’s any useful… Here is a couple screenshots from the website:

Thanks again, @andersy005!

This was exactly what I was looking for, but I didn’t come across this particular one in my searches. It’s all set up in icepyx and I’ll check that the cron ran as expected on Monday!

This project was fun to have quick visuals of GitHub stats through time (it’s similar to repo-analytics without the requirement to grant any sort of access to it; I will admit I’m a fan of the xkcd-style graphs though!). It bumps up against the issue I was having with a few other tools I found, which is that it doesn’t give you the underlying data or actually help with tracking.

For anyone wanting to check out what I implemented, see the icepyx repo on GitHub and check out our documentation in the coming days (after the PR is reviewed!) for a new tracking section!

1 Like