cloud-native-geospatial

GeoZarr multiscale pyramids for planetary-scale rasters: a lazy, windowed and resumable writer of zarr v3 stores with sharding, for any xarray object with a CRS.

Kentucky's KyFromAbove program provides statewide 2ft DEMs, 3-inch orthoimagery, 3-inch oblique imagery, and COPC LiDAR point clouds — all publicly available on S3 with a STAC API for discovery. abovepy gives you Pythonic access to all of it, plus server-side terrain analysis via TiTiler. No credentials required.

FlatGeobuf Overviews (FGBO): a scale-optimized, fully fgb-compatible vector format — precomputed simplification, overview levels, and segments for fast on-demand MVT tiles from one static file. Rust encoder/decoder, CLI, tile server, and wasm browser reader.

Cloud-native geospatial related content/activities at the 2025 ESA Living Planet Symposium

FaultLine is a cloud-native geospatial change engine for comparing, validating, and tracking spatial dataset evolution across modern formats like GeoParquet, COG, and PMTiles.

Compares VKG approach against PostgreSQL + REST API for cloud-native geospatial data integration. Federates Sentinel-2 COGs (raster) and Overture Maps GeoParquet (vector) via DuckDB FDW, PostGIS, and PL/Python. Evaluates two raster integration strategies, and SPARQL (Ontop VKG) vs REST API (pg_featureserv) trade-offs. Fully Dockerised.

An analytical MCP server for SpatioTemporal Asset Catalogs. Safely orchestrates AI-driven spatial workflows and pushes compute to the edge using Rust and DuckDB.

Curated resources for geospatial AutoML, algorithm discovery, agents, datasets and benchmarks.

Validate whether a GeoTIFF is a proper Cloud-Optimized GeoTIFF and explain exactly why it is not: structural checks, HTTP range-request cost, and a derived gdal_translate fix.

OGC Building Blocks for the Portolan specification.

See your distortion. A visual-first geospatial diagnostics engine — projection x-ray, cartographic linting, spatial diffing, and autofix. Rust core, Python bindings, browser-ready via WebAssembly.

Named analysis functions over FAO FERSPAS, each served as XYZ tiles with a time axis. One file per function; nothing pre-built.

STAC dataset intelligence: find it, verify it, plan it, replay it.

A Cloud Native Geospatial study: dynamic spatial queries served entirely from the browser via DuckDB-WASM over Overture GeoParquet.

Turn a georeferenced raster time series into an interactive dataset anyone can open with a link, with the stored values intact. Static files, no tile server.

A STAC profile for describing the layers of a map: where their data comes from, how their forms and fields work, how they relate and how they are drawn. With a JavaScript package and a QGIS Server plugin.