Gowtham1105/Pan-African-Food-Security-Risk-Radar

# Pan-African Food Security Risk Radar (PAFSR) 🌍 ### *A Multivariate Machine Learning Analysis of Hunger Drivers in Sub-Saharan Africa* --- ## 📖 Executive Summary Why do some nations starve while others thrive? The **Pan-African Food Security Risk Radar** abandons static datasets to build a **live ETL pipeline** that ingests 13 years of economic, environmental, and political data from the **World Bank API**. This project investigates the hypothesis that *Food Security is primarily an agricultural problem.* Our findings disprove this, revealing that **Economic Purchasing Power (GDP)** is a 3x stronger predictor of hunger than Agricultural Production or Rainfall. Using a **Random Forest Regressor**, we successfully modeled the vulnerability of 54 African nations with **65% accuracy**, significantly outperforming traditional linear models. --- ## 🏗️ The Data Pipeline Unlike standard analysis projects that rely on pre-cleaned CSVs, this project engineers its own dataset from scratch. * **Source:** World Bank Open Data API (Live Fetch). * **Scope:** 54 African Nations (2010–2023). * **ETL Architecture:** 1. **Extract:** Python script (`wbgapi`) queries 15+ years of data for 4 distinct indicators. 2. **Transform:** Handles missing values, enforces "Wide Format" pivoting, and filters for the 54-nation African continent. 3. **Load:** Stores clean, normalized data in a local **SQLite** warehouse (`food_security.db`). --- ## 🔍 Methodology: The "Three Pillars" Model We analyzed food insecurity through three distinct lenses: | Pillar | Metric Used | Hypothesis | | :--- | :--- | :--- | | **1. Economic Access** | **GDP per Capita (PPP)** | *Can people afford food?* | | **2. Market Stability** | **Food Price Inflation** | *Are prices volatile?* | | **3. Political Stability** | **governance Estimate** | *Is conflict disrupting supply?* | | **4. Environmental** | **Food Production Index** | *Did the harvest fail?* | --- ## 📊 Key Findings & Insights ### 1. The "P …