#27 Fedora User Activity Statistics Service (FUAS)
Opened a year ago by amoloney. Modified 8 months ago

New initiative: Fedora User Activity Stats Service (FUAS)

What is this initiative about?

Develop a service that captures the user activity in the Fedora project in order to support the Fedora Strategy 2028 Guiding Star: double the number of contributors who are active every week.

Why this initiative?

In order to make this strategy goal meaningful, we need a measurement.

There are secondary insights which we could also gather, helping us towards more specific goals or identifying particular problems.

Definition of success

What is the minimal outcome you would like to see from this initiative to be satisfied?

  • A collector (or group of closely-related collectors) that dig into the Fedora Message Bus history
  • Filters for bots and spam accounts
  • A final filter to anonymize user account names
  • Datastore as a sqlite db updated weekly
  • No need for reports

What are your nice or really nice to have wishes?

  • A collector that reads meetbot logs
  • A filter that maps IRC addresses to FAS accounts (only count those that we can identify!)
  • A Jupyter notebook demonstrating some reports
  • Reports generated automatically every week

Area/community impacted

Fedora

Will this initiative impact CentOS, Fedora? All users? All contributors? A group of contributors?
All fedora contributors, but indirectly as the end user should not have any immediate impact.

Dependencies

Do this initiative have any dependencies?
Fedora Message bus, mailing list archives, discourse API, FAS

Skills needed?
Familiarity with tool 'Sextant' could be useful & DNF Countme
SQL
Fedora Messaging system(s)
FAS

Person who must or should be involved?
NA

Other work that should be completed prior to this initiative?
Not known at this time

Deadline

ASAP, but Flock (aug 1-4) would be a nice time to launch this work.


This sounds like a dream data science tool that I have always wanted in Fedora. A few thoughts!

  1. Python is also a useful skill because most (if not all) of our data tools, apps, and platforms are Python-based. You can get pretty far with only Python scripts to poke at the data we already have.
  2. It is definitely worth an exploration with the Red Hat OSPO with their data science tool for communities. There is a big investment in the CHAOSS Project and its subsequent tools for analyzing data, e.g. GrimoireLab and Augur. We have a relationship with the lead Augur developer and he also knows a good bit about Fedora already. RH OSPO has a hosted GrimoireLab platform available and is also building a new tool out in-house based on Augur.

I reviewed the ARC investigation and I think I was not clear about the scope of this task versus the original request by @mattdm. I think it would be helpful to get on a call and understand the vision better for this work: both from @mattdm on what the success criteria is and from @t0xic0der on the scope of this project.

The concern that arose for me is that we might be duplicating work and recreating the wheel. There are other communities that we could lean on for support on the technical implementation and there are also resources in Red Hat OSPO that we could leverage around metrics and data science work. For instance, we recently had a call with Cali Dolfi and James Kunstle in the RH OSPO, and it was that conversation that is concerning me about the scope of this initiative.

Again, I think a call to get all the stakeholders together at once is important. I think we always miss some people on a call and we might be contradicting each other accidentally. :grinning:

Cali tasked me at finding a new time for us all to meet because TIMEZONES, so this is in my court to set up.

Metadata Update from @amoloney:
- Issue tagged with: Needs Info

8 months ago

Login to comment on this ticket.

Metadata