Peek inside a release
prototypeRun SQL against a SportsDataverse release file without downloading it — DuckDB in your browser reads just the bytes it needs.
Every SportsDataverse dataset is published as files on a GitHub release. They're built for R and Python loaders, but nothing stops a browser from reading them — except that GitHub doesn't send the headers a browser needs, so this site relays just the byte ranges DuckDB asks for.
The box below runs in your browser. Edit the SQL and press Run. {{src}} stands for the
2026 NBA rosters file (about 70 KB).
espn_nba_rosters/rosters_2026.parquet
Try DESCRIBE SELECT * FROM {{src}} to see the schema, or a GROUP BY over a team column
once you know what it's called.