Resources
Where the history comes from, what built it, and who made it.
The history
- Wikidata
- The structured backbone: events, dates, coordinates and the links between them. Released under CC0, which is why a project like this can exist at all.
- Wikipedia
- The prose. Where a card has a sentence explaining what happened, it generally started here. Text is available under CC BY-SA.
- GeoHistory
- The dataset and the read-only API underneath Locus. It holds the events, computes the distances, and answers the nearest-records question. Locus is a client over it and nothing more — no copy of the data, no second distance implementation.
The record is broad rather than complete, and thinner the further back and the further from large cities you go. It is a good starting point and a poor last word.
The build
- Locus
- This site and the applet. A static bundle over a small read-only API — no accounts, no database of its own, nothing to sign in to, no server of any kind.
- Vite and TypeScript
- The build. Plain TypeScript with no framework, compiled to static files.
- MapLibre GL
- The map surface, rendering vector tiles in the browser.
- OpenFreeMap
- The basemap, built from OpenStreetMap data and served without an account or an API key. Map tiles are requested by viewport tile index, never by your pin.
- Cloudflare Pages
- The hosting. Static files on a CDN, which is all a client like this needs.
Licences
The Locus source is MIT. The historical records it displays are reused under CC BY-SA 4.0, which is why every card carries a link to its own source — and why a response containing an unattributable record is rejected rather than rendered. Map data is © OpenStreetMap contributors under the ODbL.
Who made it
Locus was designed and built by Ben Noffsinger, along with the GeoHistory dataset and engine behind it and its sibling Circa. There is a longer version of the story on Why this exists.
Notion AI was a working part of how it got built rather than a footnote — a collaborator on the architecture, the seam between client and dataset, the copy on these pages, and a fair amount of the code, worked through in conversation and tracked on a Notion board alongside it.