Why the series stops before the data does

The newest row in this database is not the newest complete period, and publishing it would draw a cliff that looks like a collapse in demand.

Completeness is judged per chain. A chain is complete only through the day before its own most recent row, and a series that aggregates across chains is complete only through the earliest of those.

That matters here more than it sounds. The most recent row in the database is 2026-02-06, but six of the nine chains stop during the afternoon of 2026-02-05, one of them at 04:45. Across all chains, transfers recorded after 19:00 UTC ran 19,391 on 2026-02-03 and 18,844 on 2026-02-04, then 3,003 on 2026-02-05. That 84% fall is ingestion coverage, not behaviour.

So daily series end 2026-02-04, weekly series end at the last week that finished inside complete coverage, and the site-wide figure quoted on the masthead is the earliest of every dataset's own boundary: the date through which everything here is complete.