Spatial Analysis of Childhood Vaccination Determinants in Kenya
# 🌍 Geographically Weighted Random Forest (GWRF) Dashboard
## **Spatial Analysis of Childhood Vaccination Determinants in Kenya**
đź”— **Live Dashboard:** View Here
---
## 📌 Project Overview
This project applies **Geographically Weighted Random Forest (GWRF)**, an advanced spatial machine learning technique, to uncover the determinants of childhood vaccination uptake in Kenya. Using nationally representative **Kenya Demographic and Health Survey (KDHS 2022)** data, we explored how factors such as **maternal education, poverty, ethnicity, transport access, and residence type** influence vaccination rates—while accounting for **spatial heterogeneity** across Kenya’s 47 counties.
The results are deployed in an **interactive dashboard** that allows users to:
âś… Visualize county-level disparities in vaccination coverage
âś… Explore spatial autocorrelation patterns
âś… Investigate the importance of local determinants (e.g., transport, wealth, electricity, ethnicity)
âś… Support **policy makers, public health practitioners, and researchers** in designing targeted interventions
---
## 🚀 Motivation
Vaccination remains one of the most cost-effective public health interventions, yet disparities persist in Kenya. Understanding *where* and *why* uptake is low is crucial to achieving equitable health outcomes. This project bridges **data science, geospatial analysis, and health research** to provide actionable insights that go beyond national averages—helping identify **localized barriers and opportunities**.
---
## 🛠️ Methodology
- **Data Source:** Kenya Demographic and Health Survey (2022)
- **Target Variable:** Childhood vaccination status (fully vaccinated vs. not)
- **Model:** Geographically Weighted Random Forest (GWRF)
- Captures **non-linear relationships**
- Accounts for **spatial non-stationarity**
- **Evaluation Metrics:** Mean Squared Error (MSE), R², Moran’s I (spatial dependence)
- **Deployment:** Interactive web app/dashboard
---
## 📊 Key Insights
- Significant **s …