GeoParquet
GeoParquet stores geospatial vector data — points, lines, polygons — inside standard Apache Parquet, combining columnar compression with standardized geo metadata. It is rapidly replacing shapefiles for cloud-native GIS work.
viewparquet treats GeoParquet as a first-class format. Drop a file in and it opens like any Parquet file: browse attribute columns in the grid, see the geometry column and its encoding, read the geo metadata from the file footer, and query everything with DuckDB SQL. Processing is local via DuckDB-WASM unless you explicitly add a private Cloud copy — no QGIS install or desktop GIS startup required.
Drag a .parquet or GeoParquet file into viewparquet, or stream it from S3 or a URL — handy for large open datasets like Overture Maps extracts.
The schema panel shows attribute columns and the geometry column; the metadata view exposes row groups, compression, and the GeoParquet metadata stored in the footer.
Filter features by attribute, count by category, or sample rows with full DuckDB SQL, then export results to Parquet, CSV, or JSON.
GeoParquet is Apache Parquet with a standardized way of storing geospatial vector data: geometry columns (typically WKB-encoded) plus geo metadata — CRS, geometry types, bounds — in the file footer. Any Parquet reader can open it; geo-aware tools additionally understand the geometry.
Open it in viewparquet in your browser. The file parses locally with DuckDB-WASM — no GIS software, no install — and you can inspect attributes, geometry, and metadata, and run SQL immediately.
GeoParquet geometry is usually WKB (Well-Known Binary) inside a Parquet binary column. viewparquet shows the column in the grid and its details in the schema panel, and you can query and filter rows by their attribute columns with SQL.
GeoParquet is columnar, compressed, and has no shapefile limits (no 2 GB cap, no 10-character field names, no multi-file sidecar mess). It also queries fast with engines like DuckDB and works cloud-natively over object storage.
Device files are processed client-side with DuckDB-WASM and are not uploaded unless you explicitly add a private Cloud copy. Files opened through the direct S3/URL connector stream from the source to your browser.
Drop a Parquet, GeoParquet, CSV, or JSON file into viewparquet: browse rows, run DuckDB SQL, and export results locally in your browser. A file is uploaded only if you explicitly add a private Cloud copy.