Open data
Download the datasets
Every published sport ships the same four files: three CSVs and one Excel workbook containing all of them. The point is not to show you a result but to let you inspect the inputs, reproduce the calculation, and tell us where we are wrong.
NCAA Division I FBS
College Football
Dataset 2026.2 · 157 seasons · 12,403 precomputed ranges
CSV
Championship records (CSV)
One row per championship claim: champion, name at the time, home venue, GPS coordinates, coordinate method and precision, derived town, source, and whether the claim was counted.
Download (125 KB)CSV
Epicenter ranges (CSV)
Every contiguous season range with its calculated epicenter, derived town, spread metrics, status, and algorithm version — enough to reproduce the slider.
Download (2.7 MB)CSV
Locality audit (CSV)
Ranges whose nearest-town label changed when the resolver moved to the USGS GNIS populated-places extract.
Download (1.7 MB)Spreadsheet
Complete workbook (Excel)
All three tables as worksheets in one .xlsx file, generated from the same bytes as the CSVs.
Download (2.2 MB)National Football League
NFL
Dataset 1.3 · 106 seasons · 5,671 precomputed ranges
CSV
Championship records (CSV)
One row per championship claim: champion, name at the time, home venue, GPS coordinates, coordinate method and precision, derived town, source, and whether the claim was counted.
Download (65 KB)CSV
Epicenter ranges (CSV)
Every contiguous season range with its calculated epicenter, derived town, spread metrics, status, and algorithm version — enough to reproduce the slider.
Download (1.2 MB)CSV
Locality audit (CSV)
Ranges whose nearest-town label changed when the resolver moved to the USGS GNIS populated-places extract.
Download (767 KB)Spreadsheet
Complete workbook (Excel)
All three tables as worksheets in one .xlsx file, generated from the same bytes as the CSVs.
Download (1.0 MB)National Basketball Association
NBA
Dataset 1.1 · 79 seasons · 3,160 precomputed ranges
CSV
Championship records (CSV)
One row per championship claim: champion, name at the time, home venue, GPS coordinates, coordinate method and precision, derived town, source, and whether the claim was counted.
Download (43 KB)CSV
Epicenter ranges (CSV)
Every contiguous season range with its calculated epicenter, derived town, spread metrics, status, and algorithm version — enough to reproduce the slider.
Download (714 KB)CSV
Locality audit (CSV)
Ranges whose nearest-town label changed when the resolver moved to the USGS GNIS populated-places extract.
Download (444 KB)Spreadsheet
Complete workbook (Excel)
All three tables as worksheets in one .xlsx file, generated from the same bytes as the CSVs.
Download (594 KB)Major League Baseball
MLB
Dataset 1.5 · 123 seasons · 7,626 precomputed ranges
CSV
Championship records (CSV)
One row per championship claim: champion, name at the time, home venue, GPS coordinates, coordinate method and precision, derived town, source, and whether the claim was counted.
Download (69 KB)CSV
Epicenter ranges (CSV)
Every contiguous season range with its calculated epicenter, derived town, spread metrics, status, and algorithm version — enough to reproduce the slider.
Download (1.7 MB)CSV
Locality audit (CSV)
Ranges whose nearest-town label changed when the resolver moved to the USGS GNIS populated-places extract.
Download (1.1 MB)Spreadsheet
Complete workbook (Excel)
All three tables as worksheets in one .xlsx file, generated from the same bytes as the CSVs.
Download (1.3 MB)What is in the files
The championship file is the best starting point. Each row carries the champion under the name it held that season, the home venue that season, the coordinate with its method and stated precision, the derived nearest town, the source, and whether the claim was counted toward the epicenter. Claims that are recorded but deliberately not counted — the AFL championships in the NFL file, no-champion seasons in the MLB file — are visible rather than dropped.
The ranges file contains one row for every contiguous span of seasons, so any figure the slider shows can be checked without running our code. The locality audit shows which nearest-town labels changed when the place-name resolver was upgraded.
Reusing the data
Every file is versioned and content-hashed, and the manifest for each sport lists the SHA-256 of each artifact so you can confirm you have the release you think you have. Corrections are not silent edits: the canonical data changes, ranges are recalculated, QA reruns, and a new release is published.