uweraliliane/RwandaHiddenHungerRwanda

Data-driven analytical framework to identify and address hidden hunger in Rwanda using the 2024 CFSVA dataset — includes code, geospatial models, and policy insights. # 🌍 Ending Hidden Hunger in Rwanda 📊 **NISR Big Data Hackathon 2025 — Track 2** **Team:** UWERA Liliane **Language:** Python **Platform:** Streamlit --- ## 📌 Introduction Despite notable progress, over **33% of Rwandan children under five** remain stunted, primarily due to **micronutrient deficiencies** and **chronic undernutrition**. This project leverages **CFSVA 2024 data** and advanced machine learning to: - Identify geographic **malnutrition hotspots** - Model **key risk factors** - Simulate **household-level stunting risks** - Recommend **targeted interventions** - Provide an interactive **Streamlit dashboard** for decision-makers --- 📅 **Submission Date:** 10 October 2025 ## 🎯 Objectives This project addresses **Track 2: Ending Hidden Hunger**, aiming to: 1. **Map** stunting, wasting, and underweight prevalence across Rwanda 2. **Model** malnutrition risk with ML (XGBoost, LightGBM, Logistic Regression, CatBoost) 3. **Identify** high-impact predictors of stunting 4. **Simulate** stunting probability at household level 5. **Recommend** data-driven policy & program interventions 6. **Visualize** key findings in a user-friendly dashboard --- ## 📊 Data Sources - **Primary Dataset:** CFSVA 2024 – Comprehensive Food Security and Vulnerability Analysis - **Geospatial Data:** - GADM Rwanda Shapefiles - SimpleMaps Rwanda GeoJSON --- ## 📈 Key Findings – Hotspots & Risk Factors ### 🔺 Top 5 High-Stunting Districts 1. Nyabihu 2. Rubavu 3. Rutsiro 4. Burera 5. Gakenke ### 💡 Most Predictive Risk Factors - **Mid-Upper Arm Circumference (MUAC)** - **Vitamin A intake** - **Wealth Index** - **Women's Dietary Diversity** - **Unsafe Water Source** - **Recent Illness (Diarrhea/Fever)** --- ## 🧩 Root Cause Analysis & Interventions | **Factor** | **Root Cause** | **Recommended Intervention** | |-----------------------|---------------------------|------------------------------------------------------| | Low MUAC …