jobgichuki/South-Sudan-nutrition-crisis-forecasting.

Code and analysis for a master’s dissertation predicting nutrition crises in South Sudan using Logistic Regression and Random Forest models. Combines climate, conflict, and market data to explore data-driven early warning for humanitarian planning. # Forecasting Nutrition Crises Using Climate, Conflict, and Market Data – South Sudan This repository contains the code, figures, and supporting materials for a master’s dissertation submitted to the University of East London (October 2025). The project explores the use of Logistic Regression and Random Forest models to predict nutrition crises in South Sudan based on climate, conflict, and market indicators. ## Overview The goal of this study was to test whether data-driven models can identify early signals of nutrition crises. Results showed moderate predictive performance, with Logistic Regression offering better balance and interpretability than Random Forest. The study highlights the potential of predictive analytics in humanitarian early warning systems, while emphasizing the need for improved data quality and richer feature sets. ## Contents - **/data/** – Processed datasets and metadata (restricted data not included) - **/scripts/** – Python scripts for data cleaning, modeling, and visualization ## How to Use 1. Clone this repository 2. Install required packages 3. Run scripts in `/scripts/` sequentially for data preparation and model evaluation ## Citation If you use this code, please cite: > Gichuki, J. (2025). *Forecasting Nutrition Crises Using Climate, Conflict, and Market Data – South Sudan.* University of East London. ## Contact For inquiries or collaboration: 📧 **job.gichuki@gmail.com**