tesdem/CDR-Generation-Africa---V2

# African CDR Generation Model for ML Training and SDG Prediction ## 📊 Overview This project generates **synthetic Call Detail Records (CDRs)** for all 54 African countries, optimized for machine learning training and Sustainable Development Goals (SDG) indicator prediction. The model creates realistic telecommunications data while preserving privacy and enabling development research. ### 🎯 Key Features - **🌍 Pan-African Coverage**: All 54 African countries with realistic country-specific distributions - **📈 80+ Features**: Comprehensive CDR attributes including temporal, spatial, and behavioral patterns - **🎯 SDG Indicators**: 7 SDG proxies for supervised learning (Goals 1, 5, 8, 9, 10, 17) - **⚡ High Performance**: 1M records in ~7.5 minutes with real-time progress tracking - **📦 Multiple Formats**: CSV output with optimized dtypes for ML training - **🔍 ML-Ready**: Train/validation/test splits, feature importance analysis ## 🚀 Quick Start ### Installation ```bash # Clone repository git clone github.com cd african-cdr-generator # Install dependencies pip install -r requirements.txt