Population and Access
Who each facility actually serves, and whether those people can realistically reach it. Population from WorldPop 2026 constrained at 100 m; travel time from a modelled cost-of-access surface.
Full accessibility analysis for … is being prepared.
The travel-time + Bayesian catchment pipeline currently runs only for Tanzania. The same methodology (modelled 100 m travel-time surface from the MAP 2015 friction raster, constrained WorldPop 2026 population, and a Bayesian gravity catchment model with travel-time-decayed attendance probability) will roll out to the other seven countries as country-specific inputs become available.
In the meantime, per-country facility distribution and coverage are visible in Registry Explorer.
Travel time to nearest facility
Modelled minutes-to-nearest-facility from every 300 m pixel, classified into four bands. Fast on the green end, slow on the red end.
National access breakdown
Worst-served districts (pop > 60 min)
Estimated facility catchment — Tanzania mainland (26 regions)
Bayesian gravity model (PATH-Global-Health/catchment): probability of attending each facility is inversely proportional to squared travel time (10–300 min band), fitted with an INLA spatial random effect over the friction-derived travel-time matrix. Each facility gets an estimated catchment population with a 95% credible interval. Circle size = catchment magnitude; color = quintile band.