[https://docs.google.com/presentation/d/1JA2JX7q7UUHnKlITNteM68gL-rd0_tSMhT7zX9tMueE/edit?usp=sharing](https://docs.google.com/presentation/d/1JA2JX7q7UUHnKlITNteM68gL-rd0_tSMhT7zX9tMueE/preview?usp=sharing)

Best practice for creating GeoParquet files

https://github.com/opengeospatial/geoparquet/blob/main/format-specs/distributing-geoparquet.md#best-practices-for-distributing-geoparquet see the TL;DR:

• Use zstd for compression, and set the compression level to 15.
• Be sure to include the bbox covering, and use GeoParquet version 1.1.
• Spatially order the data within the file.
• Set the maximum row group size between 50,000 and 150,000 per row.
• If the data is larger than ~2 gigabytes consider spatially partitioning the file
• Use STAC Metadata metadata to describe the data.

The demo I used

so that you can read/write the low-level details of a parquet file is with PyArrow

• https://github.com/jorisvandenbossche/2025-cloudscaping-geoparquet-workshop

Other links

• STAC & GeoParquet: https://bsky.app/profile/developmentseed.org/post/3lygffozqv22y
• GeoParquet 2.0 vs Parquet: https://rednegra.net/blog/20250925-parquet-with-geometry-type-is-not-geoparquet/ + https://cloudnativegeo.org/blog/2025/02/geoparquet-2.0-going-native/
• Nested columns are possible, see “bbox”
• Arrow vs Parquet, see also https://geoarrow.org/format.html
• 🎧 A podcast about GeoParquet: https://mapscaping.com/podcast/geoparquet-for-beginners/

<aside> 🏋🏼‍♀️

A small (formative) exercice

Construct a GeoParquet file of all the buildings in NL containing only a few attributes, and check how many there are + what is the average height/area of them.

With DuckDB.

⇒ Why 11.8M buildings while there are ~11M in the BAG?

</aside>