GFDRR/sirs-mapillary

Mapillary street-level coverage of schools (SIRS West Africa) # sirs-mapillary **Live app: gfdrr.github.io (school + coverage layers load from open data; Mapillary photo layers go through a token proxy - see `DEPLOY.md`) A lightweight web map showing where **Mapillary street-level photos** already exist near schools, for the School Infrastructure Risk Screening (SIRS) work across five West African countries: Niger, Mali, Guinea, Benin, Ghana. The goal is to assess whether free, existing street-level imagery can help read building characteristics (roof, height, structural type) from the ground - before commissioning any new photo collection. ## What it shows - **Coverage**: how much of each country's school network has nearby Mapillary photos, and how recent they are. - **Photos**: click a covered school to see the most useful nearby images (camera-facing-the-school first), and open a panoramic viewer. School locations shown are from **OpenStreetMap** (open data, ODbL). Street-level imagery is from **Mapillary**. ## Run locally It is a single static page - no build step. ```sh # Optional: for the photo layers, provide a Mapillary token cp .env.example .env # then add your MAPILLARY_TOKEN python3 -m http.server 8765 # open localhost ``` ## Deploy (public) For a public deployment the Mapillary token must not sit in the browser. A small Cloudflare Worker proxy holds the token server-side. See **`DEPLOY.md`** for the full steps (deploy the worker, set the proxy base, publish to GitHub Pages). ## Data + licenses - School points: © OpenStreetMap contributors, **ODbL 1.0**. - Street-level imagery: **Mapillary** (CC-BY-SA). - Coverage figures in `data/coverage_summary.json` are aggregate, country-level only. ## AI-assisted development > This project was developed with significant assistance from AI coding tools. - **Claude Code** (Anthropic) - code generation, architecture, and documentation. - All functionality has been tested and verified to work as intended. - Features and in …