Tracking community health via metrics

webchick's picture

Dalin pointed me to this bad-ass blog post about how Mozilla does metrics around community management:

Check this dashboard out!

Only local images are allowed.

I think something like this could be totally awesome for figuring out how we're doing in terms of onboarding new contributors, retaining existing contribitors, and gauging the overall "health" of the community.

Anyone have any other pointers to cool stuff projects are doing to make sure their contributors are healthy and happy?



arshadcn's picture

Anyone have any other pointers to cool stuff projects are doing to make sure their contributors are healthy and happy?

Ohloh,, they have some cool metrics here. I'm not sure how often this is updated though.

That blog post is great.

catch's picture

That blog post is great. Tracking activity of individual contributors, while a bit freaky, is only going to be pulling out publicly available data (commits, issue followups etc). although still needs to be handled carefully.

I can't provide examples of what other projects do, but just to note there's an older post in this group discussing possible additional metrics to measure for Cross-posting just in case. Those are separate discussions to have (reviewing other projects, figuring out what's desirable and feasible on, and tracking aggregate vs. individual contributions) so it makes sense to keep them separate.

marvil07's picture

Some time ago I started trying to build a way to know who is contributing to core, and the end up is on a sandbox, that use code history( in general public-easy-to-access-available-information) to generate stats, and I really need to improve them(have plans to integrate it with more external tools that analyses code).

If we are running anything on, we naturally can provide a lot more than that, but the problem actually is to define what aspects to track(and this group is also about defining that, I guess ;-) ).

I completely agree that building more tools on about how to track community health is a big win, but IMHO exposing the data to let people build better things around tracking contribution will be the best first step. For example: Project API, a way to access issue queue information programatically(REST too?), and other relevant data that could provide information about contribution.

Just for the record the data on the picture is basically commits(seems like they assume 1 patch= 1 commit), so it should be totally possible to get that information now(we have that info already on the database), but that leaves out many contributors(i.e. patch reviewers).

yes please!

leisareichelt's picture

so, if we could have a dashboard kind of like that, that tracked the metrics we've got listed here:
that would be super awesome.

i wonder where it would live.

leisa reichelt -


webchick's picture

Thanks a lot for the pointer to that other thread.

I think I can gather a fair number of these with some SQL queries against the database. I'll do some futzing this week.


catch's picture

More issue queue metrics we could look at:

Average number of replies to an issue (this is not necessarily a good or bad thing, although too many replies is usually bad). This could be split between open and closed issues. Subscribe/+1 is going to skew this a bit, but those are also real followups that bump issues to the top of people's trackers so this is possibly fine.

Distribution of number of replies - i.e. how many issues have 0, 1, 2, 3, 4, 5, 10, 20, 50, 100 replies (again open vs. closed).

All of this could be run against the issue queue as a whole, just for core, just for some specific contrib projects (top 100 usage or whatever).

Number of issues total over time (closed vs. open again) - would make sense to split this by bug report / task etc.


David Eaves Image Is Here