zeroin123/EHAI-Project

Ethiopian Health access index project. Assessing the operational status of Ethiopia's Health facilities at the woreda(district) level using VIIRS satellite dataset as a proxy for their electrification # EHAI Ethiopia — Effective Healthcare Access Index > Distinguishing *paper coverage* (a facility exists) from *effective access* (the facility is electrified and operational) across Ethiopia's 1,148 woredas. Built for the **HabTech Hackathon 2025** using VIIRS 2025 nighttime light satellite data, the Ethiopia MoH HFRHIS facility registry (~40,000 facilities), and WorldPop 2020 population estimates. --- ## Key Findings | Metric | Value | |---|---| | Dark facilities (likely non-electrified) | **26,019 of 40,478** (64.3%) | | Effective healthcare coverage | **7.4%** of the population | | Population at risk | **107.3 million people** | | Paper-coverage woredas | **1,026 woredas** — high nominal but low effective access | | Mean EHAI score | **18.1 / 100** across 1,148 woredas | --- ## Quick Start — Dashboard Only (no pipeline needed) Pre-computed outputs are included in this repository. The dashboard runs immediately after install. ```bash git clone cd healthcare-access-index pip install -r requirements.txt streamlit run dashboard/app.py ``` Open localhost in your browser. > **Python version note**: The geospatial stack (rasterio, geopandas, rasterstats) requires Python 3.11. > If your default `python` is newer, use the full path: > ``` > C:\Users\ \AppData\Local\Programs\Python\Python311\python.exe -m streamlit run dashboard/app.py > ``` --- ## Full Pipeline Reproduction To re-run the full analysis from raw data (re-extracts NTL values, recalculates EHAI scores): ### Step 1 — Download VIIRS NTL (manual, one-time, free account required) The VIIRS raw file is ~11.6 GB and cannot be hosted on GitHub. 1. Go to **eogdata.mines.edu 2. Create a free account (Colorado School of Mines / NOAA EOG) 3. Navigate to: Annual VNL V2 → 2025 → vcmslcfg → average_masked 4. Download the global composite `.tif.gz` 5. Decompress and save to: ``` data/raw/ethiopia_ntl_viirs.tif ``` ### Step 2 — Download all other datasets (automated) `` …