Data Explorer works offline

Something that surprises people: load Data Explorer once, then turn off your network. It keeps working.

There is no API behind it. The page ships the code, the datasets are static files, and the computation happens in your tab. The only thing that needs the network is fetching a dataset you have not loaded yet.

That is a side effect of the no-server decision rather than a feature we built, but it is a real one — this stuff works on a plane, and it works behind a corporate proxy that blocks everything interesting.

It has a longer-term version too. A site with no backend has nothing to keep running: no database to migrate, no API to version, no service to pay for. The failure mode of an abandoned static site is that it keeps working, which is not true of anything with a server behind it. Most of the data tutorials that were useful ten years ago are gone, and almost none of them died of being wrong.

Static sites are underrated. The whole thing is files, and files last.

Data Explorer — A fast table with per-column profiling: type inference, missing counts, distinct values, quantiles, and a sparkline distribution for every numeric column.

All seven tools are on the toolkit page, and every one of them reads the same dataset library.