Skip to main content

Blazing Fast Geospatial SQL in DuckDB

· One min read
Isaac Brodsky
Isaac Brodsky
CTO @ Fused

In this video from the FOSS4G 2024 conference, Isaac Brodsky, CTO and co-founder of Fused, shows the power of combining H3 with DuckDB to enhance geospatial data analysis.


As an example, Isaac shows a Fused User Defined Function (UDF) that joins the Overture Places dataset and the Natura 2000 biodiversity areas dataset, achieving significant reductions in file size and query execution times. He showcases how the integration allows for efficient data exploration, filtering, and real-time queries, emphasizing the power of DuckDB's H3 extension and spatial extension.

Isaac explains how H3 simplifies geospatial analytics by offering common spatial index to join datasets and enables efficient storage & processing by converting spatial features into 64-bit integers. Additionally, DuckDB enables developers to conveniently transition between Python and SQL. He also highlights how DuckDB can simplify data processing architectures by querying data in real-time from object storage systems such as S3.