Work

Pipelines, tools and open data. Each write-up covers what the work does, how it was built and what it produced.

Deaf community places, worldwide

3,422 schools, clubs, associations, sports clubs, interpreting services and places of worship in 124 countries, found in Overture Maps and checked against Wikidata, OpenStreetMap and Deaflympics member lists. The map reads vector tiles straight from a file; there is no map server. How it was built

Open data

Deaf community places
A global open dataset built from Overture Maps, corroborated against Wikidata, OpenStreetMap and Deaflympics member lists, published as GeoParquet and PMTiles.

Tools

poi-community-mapper
The engine behind the Deaf places dataset: profile-driven extraction, multilingual name classification, authority matching and GeoParquet publishing for any community. Python, DuckDB, geoparquet-io, tylertoo.