sqlrooms-flowmap, a Python tool that uses DuckDB with spatial extensions to prepare OD data for tiled serving:

• Hierarchical clustering: locations are grouped at each zoom level using pixel-radius clustering, creating a hierarchy where clusters merge as users zoom out
• Nested Hilbert indexing: OD pairs are indexed using a space-filling curve that preserves locality, enabling efficient range queries for tile-based serving
• Spatio-temporal aggregation: flows are aggregated to match zoom-based clustering, with optional temporal bucketing by hour/day/week
• Output format: Parquet files ordered by (zoom, hilbert_index) for sequential reads and efficient tile extraction
•  Browser visualization: SQLRooms + flowmap.gl renders clustered flows with smooth zoom transitions