lukanora/nigeria-food-insecurity

Analyses malnutrition and food insecurity across Nigeria's 6 geopolitical zones, then builds 3 machine learning models to predict which households face severe food insecurity with the best model reaching ROC-AUC of 0.970. # 🌾 Predicting Severe Food Insecurity & Malnutrition in Nigeria **Author:** Onyelukachukwu Nora Gwam **Tools:** Python · pandas · scikit-learn · matplotlib · seaborn **Data:** Nigeria Household Food Security & Nutrition Indicators — 900 records across 35 states and FCT (NBS/UNICEF/WFP-inspired, 2023) **Best Model:** Logistic Regression — ROC-AUC: **0.970** --- ## 🎯 Project Overview Nigeria is home to over 25 million food-insecure people, with the worst crisis concentrated in the North East and North West — regions affected by conflict, poor rainfall, limited market access, and low dietary diversity. Children under five in these areas face wasting rates above the WHO emergency threshold of 10%, with long-term consequences for cognitive development, productivity, and economic growth. This project applies **exploratory data analysis and machine learning** to 900 simulated household-level records to: - Map malnutrition and food insecurity across Nigeria's six geopolitical zones - Identify the structural and agricultural drivers of severe food insecurity - Build predictive models that could support early warning systems for government and NGO food security programmes --- ## 🔍 Key Questions Answered 1. How does child malnutrition (wasting, stunting, underweight) vary across Nigeria's six geopolitical zones? 2. What is the relationship between conflict exposure, dietary diversity, rainfall, and food insecurity? 3. Which machine learning model best predicts severe household food insecurity? 4. What are the most important drivers of severe food insecurity — and what do they imply for policy? --- ## 📊 Visualisations Produced | Figure | Description | |--------|-------------| | `fig1_malnutrition_by_zone.png` | Wasting, stunting, and underweight rates by geopolitical zone vs WHO emergency threshold | | `fig2_diversity_conflict_scatter.png` | Dietary diversity vs food insecurity + conflict exposure vs child wasting (scatter plots with trend lines) | | `fig3_rainfall_i …