A fast native desktop app to open, inspect, fix and query GeoParquet. Opens datasets too heavy for other tools, grades every file against best practices, rewrites slow files in one click, runs spatial SQL on local and remote data, and has every dataset below built in. Pure Rust, no server, no GDAL, for macOS, Windows and Linux.
geomermaids · Projects
What I build in the open, beside client work. All open source, all cloud-native, all about making geospatial data easier to use.
The world's power grids, telecom masts, pipelines, wells and water works from OpenStreetMap, measured against authoritative open sources (so far the EIA for US power plants and the NSTA for UK oil and gas) and completed where OpenStreetMap falls short. Every row says where it comes from, and every weekly build publishes how OpenStreetMap compares with each source. 27 layers, one GeoParquet file each, ODbL.
Free GeoParquet 2.0 you query straight from a URL: daily OpenStreetMap for the US, Canada and Mexico (98 regions, 16 themes), GMWID, the world's infrastructure (power, telecoms, oil and gas, water) from OpenStreetMap enriched with authoritative sources, FAO GAUL 2024 administrative units worldwide, Corine Land Cover 2018 for Europe, and FEMA flood zones for the US. STAC catalogs listed in the Portolan registry. Built by an open-source pipeline from Geofabrik's OSM extracts.
Experiments
Ideas tried out in code, to read and reuse.
An async Sentinel-2 tile server: renders scenes as XYZ tiles on demand, reading Cloud Optimized GeoTIFFs directly from object storage, with no pre-rendering and no database.
On-demand climate risk indicators from Sentinel-2 imagery (heat, drought, flood and vegetation stress) for any parcel in the world, designed for non-specialist decision makers.
Working on something geospatial and want a second opinion, an architecture review, or a hands-on contributor? Get in touch.