crayglockes/kenya-healthcare-inequality

A quantitative spatial analysis of healthcare access inequality across Kenya's 47 counties, applying econometric methods from applied statistics. # 🏥 Kenya Healthcare Access Inequality Study > A quantitative spatial analysis of healthcare access inequality across > Kenya's 47 counties, applying econometric methods from applied statistics. **🗺️ Interactive Map →** ## Research Question Is there systematic spatial and socioeconomic inequality in healthcare access across Kenyan counties, and what socioeconomic factors predict it? ## Methodology **Data Sources:** KNBS Census 2019, KHIS 2022, KIHBS 2021/22 Poverty Estimates, UNDP Human Development Index 2022, KeNHA Road Density 2022 **Analytical Framework:** 1. **Inequality Measurement** — Gini coefficient, Theil T index (between/within decomposition), Lorenz curves 2. **Spatial Analysis** — Global Moran's I, LISA cluster mapping (Queen contiguity, 999 permutations) 3. **Regression Modelling** — OLS log-log specification, interaction model for urban-rural moderation; HC3 robust standard errors 4. **Robustness** — Breusch-Pagan test, VIF multicollinearity analysis, jackknife LOO stability ## Key Findings | Hypothesis | Result | p-value | |---|---|---| | H₀₁: No spatial autocorrelation | **Rejected** | < 0.01 | | H₀₂: Poverty has no effect on access | **Rejected** | < 0.001 | | H₀₃: No urban-rural moderation | See report | — | **Gini Coefficient (HFD):** ~0.42 — comparable to sub-Saharan income inequality **LISA Analysis:** Clear deprivation cluster in North Eastern region **Theil Decomposition:** ~60% of inequality attributable to between-region variation ## Skills Demonstrated `Spatial Statistics` `Inequality Measurement` `Econometrics` `GeoPandas` `Statsmodels` `Moran's I` `LISA` `OLS` `Robustness Checks` `Folium`