Page MenuHomePhabricator

[SPIKE] Investigate if there are particular categories of image that get deleted more often
Open, Needs TriagePublic

Description

As part of the new designs for the "describe" step we were thinking of add a "what class of image is this?" question with options like "people", "nature", "buildings", etc

Users are already adding categories to their uploads, so perhaps we can use the categories they're already adding instead

The reason for doing this is we want to spot uploads that are more likely to be copyvios earlier in the process - either directly in UW itself, or by making them easier for admins to find

So a good first step would be to dig into historical DRs and deletions-without-DRs, and see if particular categories tend to cause problems. This won't be trivial, as categorylinks are not preserved for archived File pages, but we should be able to retrieve the data