README.md
In ansible-community/stats-map: Ansible Contributor Map

Ansible Map

This is a remake of the map tab currently found on https://stats.eng.ansible.com/apps/meetups (second tab)

It is a work-in-progress and will eventually replace the current map.

The map uses two data sources

1) The Meetup API data generated from the current meetup app a) Eventually, this needs to move to here or to a new backend project 2) The PR data generated by the Collections crawler b) This lives on Github too

1) Meetup

We can do quite a bit with the meetup data. Ideally we want to represent location, size, and activeness of each meetup, which we do like this:

Location - Meetup API will provide a lat/lon of the meetup, which we use
Size - This is a bit arbitrary, as we don't really know how many people actually turn up. So, we use the size of the Meetup group
Activeness - Harder, see below

For activeness, we really care about the combined effect of a few things. We know some things about all the events the group has had to-date, such as when it was and how many people RSVPd. That means we can calculate:

The frequency (n/year)
The reliability of the scheduling (the variance of the gaps between events)
Attendance trend (modelling RSVPs and seeing what the slope is)

Combined in some way (TBD) these determine an "activity" which we can use as a transpancy layer on the map, effectively hiding inactive meetups (perhaps a floor of 0.1 so they can still be just-about seen?)

2) GitHub

The Stats Collections Crawler runs nightly and indexes the Issues and PRs of every collection to MongoDB. We can use this as follows:

Get a unique list of issue/PR authors across all the collections
Peform GH API queries for their location data (this is an arbitrary string)
Perform Geocoding on the location strings, discarding anything that doesn't work (null, localhost, behind you, etc)
Just keep the "country" field from the geocoding, for privacy
Join this back to the Issue/PR data

At this point we can do a count of PRs/country and use that to colour the map.

I'm assuming that people who put real location data in a GH profile don't mind people looking at it - but this is a grey area. If you do not want to be included in the map, mail me at gsutclif AT redhat DOT com, and I'll mark your username for exclusion.