[15:04:53] danilo: what are you thinking as far as uses for the data? [15:17:42] isaacj: my initial motivation to do that was the Movement Strategy "Identify Topics for Impact" recommendation, it can also be used as an complement to [[meta:List of articles every Wikipedia should have]], to prioritize those articles in small wikis [15:21:41] theoretically, if an article is popular in 10 big wikis, the people who speek languages of small wikis can also be interested in the same topic [15:22:15] yeah, makes sense. i know some of the existing recommender systems on wiki and campaign worklists use pageviews as a metric for prioritizing content. it's interesting to see what sorts of articles have broad interest vs. more specialized interest in your data. there was some work a few years back too that you might be interested in around trying to predict pageviews in an article if it's translated into a language: [15:22:15] https://arxiv.org/abs/1604.03235. Nettrom has some really great work from a number of years back on this topic of misalignment too (quality vs. pageviews) that I often think about: https://www-users.cse.umn.edu/~bhecht/publications/wikipedia_supplydemandquality_icwsm2015.pdf [15:54:27] that research about quality and pageviews misalignment is interesting, I came to similar conclusion in other works I made related to articles prioritization, I said some about that in this topic: https://meta.wikimedia.org/wiki/Talk:Movement_Strategy/Initiatives/Identify_Wikimedia%27s_Impact#Experiment_in_Portuguese_Wikipedia [15:57:46] one of the ways I found to visualize that relation between pageviews and quality is the 2d historam of pageviews (y axix) X page size (x axis) at the end of this page: https://ptwikis.toolforge.org/Prioridade [15:58:21] I am planning to make that graph for all wikis [16:01:17] Hi all! Friendly reminder that the research showcase will start in about 30 minutes!