AndrewTavis_WMDE added a comment.
@Manuel, something else to explore would be to see if we could figure out a
metric that links `user_agent` and `ip`. I'm a bit confused why we'd go through
this and then still we have tons of unique individuals within Python's requests
count. A breakdown:
- User has unique `user_agent` and `ip`
- All's fine - select the `user_agent`/`ip` pair
- User has unique `user_agent` and multiple `ip` values
- We're assuming that the `ip` value was changed by their provider and
select one `user_agent`/`ip` pair
- User has a unique `ip` but is using multiple `user_agent` values
- We select one of the `user_agent`/`ip` pairs assuming that it's unique
and that the others are generated repeats
This basically boils down to:
- A selection of `user_agent`/`ip` pairs where each value can only occur once
within the pairs
For the above we'd get a few extra values based on those users who are
generating and getting their `ip` switched, but then this might be better than
dramatically undershooting our total user base? Yes it changes the metric, but
the goal is an estimation of usage that the Python requests `user_agent` is
masking.
TASK DETAIL
https://phabricator.wikimedia.org/T334558
EMAIL PREFERENCES
https://phabricator.wikimedia.org/settings/panel/emailpreferences/
To: AndrewTavis_WMDE
Cc: mforns, xcollazo, Ottomata, lbowmaker, WMDE-leszek, AndrewTavis_WMDE,
Michael, Manuel, Astuthiodit_1, karapayneWMDE, Invadibot, maantietaja,
ItamarWMDE, Akuckartz, Nandana, Lahi, Gq86, GoranSMilovanovic, QZanden,
LawExplorer, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, Mbch331
_______________________________________________
Wikidata-bugs mailing list -- wikidata-bugs@lists.wikimedia.org
To unsubscribe send an email to wikidata-bugs-le...@lists.wikimedia.org
