RapidAce/humanitarian-aid-africa

# 🌍 Data-Driven Humanitarian Aid Allocation Across Africa **An end-to-end data analysis and machine learning project for optimizing humanitarian resource distribution across Direct Aid Society's 30-country operational footprint.** --- ## 📋 Project Overview Direct Aid Society is a Kuwait-based humanitarian NPO operating across 30+ countries in Africa and Yemen, focused on education, water & health, economic empowerment, and emergency relief. This project analyzes 15 publicly available development indicators across 30 countries (2010–2023) to answer a core operational question: **where should humanitarian resources be allocated for maximum impact?** ### Key Findings | Finding | Evidence | |---------|----------| | Central African Republic is the **only country in active crisis** | High need AND deteriorating — the sole URGENT quadrant country | | Education investment **reduces child mortality** | Statistically proven: p=0.032, Cohen's d=0.94 (large effect) | | Food insecurity is **spreading** | 5 countries worsened by >5 percentage points since 2018 | | Water access is the **#2 predictor** of future deterioration | Feature importance from Random Forest analysis | | 14 countries are improving | SUSTAIN quadrant — programs are working, maintain funding | --- ## 🏗️ Project Structure ``` humanitarian-aid-africa/ │ ├── data/ │ ├── raw/ # Raw API pulls │ └── processed/ │ ├── master_dataset.csv # Clean master dataset (420 rows × 20 cols) │ ├── humanitarian_aid.db # SQLite database │ ├── country_trends.csv # Trend analysis per country × indicator │ ├── priority_matrix.csv # Aid priority quadrant assignments │ ├── scaler_params.json # ML model: StandardScaler parameters │ ├── cluster_centroids.json # ML model: K-Means centroids │ ├── pca_components.json # ML model: PCA transformation matrix │ └── feature_ranges.json # ML model: indicator min …