[14:08:53] joal: currently looking at https://wikitech.wikimedia.org/wiki/User:Joal/WDQS_Traffic_Analysis Super interesting. And thanks for adding all the explanations. Very helpful. [14:12:31] Lydia_WMDE: Thanks for reading :) I'm willing to write a similar page for query analysis (parsing queries and analyzing their content) [14:12:49] that'd be awesome [14:13:04] and as Léa said also something we should show the community [16:00:34] isaacj: thanks for sharing this. I slept on this particular question (about where to link back the projects for which we create hub meta pages) and I'm thinking instead of "A project by Wikimedia Research" and linking to research.wikimedia.org we can say "A project in program_name" and we link the program_name to one of the 3 main programs we have in https://research.wikimedia.org/projects.html . [16:01:11] isaacj: This way, the reader can go back to the "parent" of the project in research.wikimedia.org with one jump. [16:02:13] yeah, that sounds fine to me [16:02:45] isaacj: the blurriness of brands is usually problematic when we have external folks involved. They talk with one entity and they assume another entity has "approved" their work. They then take certain actions and they receive major pushback and some of us will have to spend time on putting out fires. :) That's my main challenge. [16:03:07] isaacj: I feel names are there to create clarity and if they confuse folks, then we should change them. ;p [16:03:57] isaacj: I do like the spirit of what you say very much. I hear it as "less strict brands create space for more fluidity" and I love it. :) [16:06:43] yeah, and while I'm not strongly opposed to experimenting with this, i'm not sure that branding would actually clear up the confusion you mention. e.g., even though WMF Research for example might feel different from Wikimedia Research to us, I'm not sure that it would immediately read as different to someone who is less steeped in the different communities (who reasonably might just focus on that the titles include wiki + research). [16:06:43] of course depends what is ultimately chosen, but just my thinking at the moment [16:08:01] that said, i obviously am open to experimentation and curious to hear other viewpoints (my understanding is that this was a larger conversation before I joined WMF) [16:31:24] Has anyone been facing weird performance issues when running spark/hive jobs via jupyter notebooks? I was having some issues since Monday evening, and today Tiziano and joal from analytics also confirmed it! [16:31:37] anyway, I have raised a phab ticket https://phabricator.wikimedia.org/T258612 [18:01:49] isaacj: a heads up that I asked Deb Zierten to give you, Martin and myself comment rights on a deck about readership growth research. I did a pass over it and I see opportunities to share some of our findings and also give feedback and help improve their research (which is btw, helpful for our readership gap measurements as well). [18:14:36] leila :thumbs up: i'll take a look when it arrives