Data status

Each data refresh is checked before anything is rebuilt: columns and types, row counts, duplicates, missing values, and freshness, sport by sport.

Before any page is rebuilt, each sport’s new data goes through a set of checks: the columns and types the models rely on, row counts (a pull may not lose more than 1% of the rows it had), duplicates, missing values, and whether the data is current. If a check fails, that sport’s update stops and the site keeps showing the last data that passed. This page shows the most recent result for each sport.

Loading…

The checks are defined in ingest/, and the columns and rules each sport’s data must meet are in contracts/.