Spatial Access Measures is a Statistics Canada dataset that shows how easy it is to reach essential places like jobs, schools, healthcare, and stores using public transit, cycling, or walking. Refer to the methodology report to see how the data was created.
This project reimagines access to the SAM dataset by transforming the bulky CSVs into a lightweight, spatially-aware GeoParquet format. With this, users can run SQL queries directly in their browser—filtering, mapping, and exploring accessibility metrics for neighborhoods across Canada, all without any server backend.
See the notebooks directory for Jupyter notebooks on how CSV data was transformed into WKB encoded GeoParquet files.
The app involves:
- Cloud-hosted GeoParquet dataset
- DuckDB-WASM for in-browser SQL analytics
- Apache Arrow tables piped directly into Deck.GL for rendering
- No backend server: everything runs in the browser
Full implementation notes are available in Technical.md.
Minimum Node 23 required. Install Node using nvm.
pnpm install # install dependencies
pnpm dev # start local dev server- Statistics Canada for the Spatial Access Measures dataset
- DuckDB for enabling fast, in-browser SQL analytics via DuckDB-WASM
- Deck.GL for high-performance GPU-based geospatial visualizations
- GeoArrow for efficient binary geospatial data interchange using Arrow format
- Apache Arrow for powering the columnar data pipeline between DuckDB and Deck.GL
Licensed under MIT. Contributions welcome!