Reading Portolan Data

Explore, analyze, and visualize cloud-native geospatial data from Portolan catalogs.

A Portolan catalog is a STAC-based collection of cloud-native geospatial data — GeoParquet files for vectors, COGs for rasters, COPC for point clouds — served as static files from object storage. No API server needed. You query the data directly with DuckDB, read metadata from STAC JSON, and visualize with PMTiles.

Specification: Portolan is defined by the Portolan Spec. For the most up-to-date details on catalog structure, format requirements, versioning, and best practices, read the spec files directly — especially core.md, structure.md, versions.md, and the format-specific docs under formats/. This skill summarizes the key points, but the spec is the authoritative source.

Tools & Installation

Check which tools are available before starting. Install what's missing.

Required

DuckDB — Primary analysis engine. Handles SQL queries, spatial operations, Parquet reads, and HTTP range requests against remote files.

# Check
duckdb --version