bellonaut/health-desert-scorer

Geospatial ML for under-5 mortality risk prediction in Nigeria # Health Desert Scorer (Nigeria) A web-based planning tool to identify Local Government Areas (LGAs) with higher barriers to healthcare access. **Standalone public route:** `/ng` on the deployed FastAPI host **Legacy Streamlit fallback:** bashir-healthdesert.streaml… **Website:** bashir.bio --- ## Global Landing Routing Planned production routing: - `healthdesert.io` -> globe landing page (static) - `healthdesert.io` -> Nigeria standalone PWA - `healthdesert.io` -> Canada app (future) Implementation notes are in `docs/global_landing_deployment.md`. Redirect assets live in `app/static/_redirects`, `app/static/ng/index.html`, and `app/static/ca/index.html`. --- ## Why this tool exists Healthcare outcomes are strongly influenced by access barriers, including distance to care, facility availability, and connectivity constraints. This tool supports evidence-informed planning at LGA level. --- ## Who can use this - Frontline health teams - NGO program teams - Government policy units - Donors and implementation partners - Researchers and students --- ## Quick start 1. Open the deployed standalone dashboard at `/ng` on the public host, or use the legacy Streamlit fallback: bashir-healthdesert.streaml… 2. Select a **State** (or keep **All Nigeria**). 3. Select a **Year**. 4. Use the left panel to search an LGA and set a focus: - All risk - Child mortality - Facility access - Connectivity - 5km coverage 5. Review map layers (**Risk score**, **Facilities**, **Connectivity**) and the highest-need list. 6. Switch to **Research** view to access feature attribution and compare workflows. --- ## Key definitions - **Health desert:** area where people face stronger practical barriers to care. - **Risk score:** relative planning score (0-10), not a diagnosis. - **Highest-need list:** ranked LGAs under current filters. - **Connectivity layer:** LGA connectivity signal (towers per 10k where …