L. Mansfield

Data status

The site’s data, tested nightly.

Every night one job pulls this site’s sources, rebuilds its tables and tests them: row counts, blanks, unique keys, allowed values, totals that must add up, and how fresh each source is. This page shows the last run as it happened, failures included.

Loading the last run…

How the run works

  1. Extract. Rebuild the site’s tables from its own pages, read the fraud pipeline’s published snapshot, and pull public activity from the GitHub API. A source that can’t be reached keeps its last good data and fails its freshness check here.
  2. Test. Hold every table to the tests written down once, check every column is described, and rerun each “SQL source” query to prove it still returns what its page shows. The data catalog lists every table, column and metric those tests cover.
  3. Log. A scheduled GitHub Action commits the results, the site redeploys, and this page reads them.