The page size in bytes would be an interesting ranking signal. In the Wikimedia SQL dumps, there is a page_len attribute in the page table. We could aggregate the counts across all Wikis (perhaps excluding Wikidata), and export the average (median? 90th percentile?) as a ranking signal.
brawer
commented
5 months ago
The page size in bytes would be an interesting ranking signal. In the Wikimedia SQL dumps, there is a
page_lenattribute in thepagetable. We could aggregate the counts across all Wikis (perhaps excluding Wikidata), and export the average (median? 90th percentile?) as a ranking signal.