[Wiki-research-l] Monthly (predicted) article quality dataset

Hey folks, I just finished working with Amir[1,2] and building off of some of Morten's work[3] to put together something that I think you're going to like. Halfaker, Aaron (2016): Monthly Wikipedia article quality predictions. This dataset contains a row for every article-month since 20010101. Each row has an article quality prediction based on text-only machine classifier (from [3] with slight improvement) and hosted by ORES[4]. We've managed to build models for English, French, and Russian Wikipedia, so I've generated datasets for each of those wikis. It's current as of 2016-08-01 and I plan to run updates periodically. Here are the columns: - page_id -- The page identifier - page_title -- The title of the article (UTF-8_with_underscores) - rev_id -- The most recent revision ID at the time of assessment - timestamp -- The timestamp when the assessment was taken (YYYYMMDDHHMMSS) - prediction -- The predicted quality class ("Stub", "Start", "C", "B", "GA", "FA", ...) - weighted_sum -- The sum of prediction weights assuming indexed class ordering ("Stub" = 0, "Start" = 1, ...) I'll update the docs based on your questions :) 1. https://phabricator.wikimedia.org/p/Ladsgroup/ 2. https://github.com/Ladsgroup 3. http://www-users.cs.umn.edu/~morten/publications/ wikisym2013-tellmemore.pdf 4. https://ores.wikimedia.org/ -Aaron