Data Reference
The layers below are the inputs the spatial notebooks read. They are hosted on the
Hugging Face dataset nhic/rwanda-spatial
— not committed to this repo — and download on demand. Everything is in the WGS84
geographic coordinate system (EPSG:4326).
Administrative boundaries
Official National Institute of Statistics of Rwanda (NISR) boundaries, 2006 vintage. Provinces, districts, and sectors ship together as a single shapefile set; cells and villages are separate.
Provinces, districts, and sectors — NISR 2006, WGS84. Shapefile set with sidecars.
Sectors (admin3) are the join granularity for climate data:
Provinces (rwa_adm1.geojson) and districts (rwa_adm2.geojson) are available the same way.
Sectors (admin3) are the join granularity for climate data. The relevant attribute
columns are ADM2_EN (district) and ADM3_EN (sector) — combined into a single key for
merging. See Climate Maps.
Cells (admin4) and villages (admin5) are too granular to preview but are downloadable:
Environmental & context layers
Vector layers used to give maps physical context (water, wetland, land cover).
Lakes/waterbodies, rivers, land cover, and protected areas (polygons + lines).
After unzipping, loading any layer is the same one-liner:
import geopandas as gpd
wetland = gpd.read_file("Wetland/Wetland.shp")
wetland.plot()
Climate data
Monthly temperature (avg/min/max), relative humidity, and precipitation per health facility.
Values are located via the DHIS2 org-unit hierarchy (orgunitlevel1–orgunitlevel5). The
Climate Maps page documents how its columns are
cleaned and joined to sectors.
climate-data.xls is the legacy .xls format — read it with engine="xlrd" and
skiprows=1 to drop the two-row header.