Besides the existing Meta-wiki page that Nemo seems to prefer, it might be also worth reviewing and comparing the standard definition used for readership data (pageviews, unique devices...), which I believe is encoded at [1]. And perhaps also the conceptual work that originally went into it [2] (although I believe we since deviated from it by e.g. including chapter wikis in the pageview data).

[1] https://github.com/wikimedia/analytics-refinery-source/blob/master/refinery-core/src/main/java/org/wikimedia/analytics/refinery/core/PageviewDefinition.java#L75
[2] https://meta.wikimedia.org/wiki/Research:Page_view/Generalised_filters#Filtering_to_applicable_sites


On Tue, Jun 26, 2018 at 6:40 PM, Neil Patel Quinn <nquinn@wikimedia.org> wrote:
Hey everyone!

As you probably know, the Wikimedia cluster includes not just "normal" wikis like English Wikipedia and Albanian Wiktionary, but odd ones like the Wikimedia Belgium chapter website, Test Wikidata, and the English Wikipedia Working Group on Ethnic and Cultural Edit Wars wiki.

As far as I know, however, there's no standard definition of "normal" wiki to use when doing analysis.

So I've started meta:Research:Wiki with draft definitions of "public wikis" and "content wikis", along with some initial documentation about wiki metadata (names, project groups, etc.) which I plan to continue to work on.

I encourage you to edit or comment on the talk page!

--
Neil Patel Quinn (he/him/his)
product analyst, Wikimedia Foundation



--
Tilman Bayer
Senior Analyst
Wikimedia Foundation
IRC (Freenode): HaeB