dkoevidjin2-ops/africa-co2-spatial-analysis

Spatial analysis and prediction of CO₂ emissions in Africa using Python, GIS and Machine Learning. # 🌍 Spatial Analysis and Prediction of CO₂ Emissions in Africa ## Overview This project presents an end-to-end geospatial data science workflow for analysing and predicting CO₂ emissions across Africa using the EDGAR 2025 database. The study integrates Geographic Information Systems (GIS), spatial statistics and machine learning to identify emission hotspots and forecast future emission patterns. The project was developed to support evidence-based climate policy and sustainable development planning. --- ## Objectives - Analyse spatial distribution of CO₂ emissions. - Detect statistically significant emission clusters. - Identify hotspot and coldspot regions. - Develop predictive models for future emissions. - Build interactive dashboards for decision-makers. --- ## Dataset Source: EDGAR 2025 Variables include - CO₂ emissions - Country - Coordinates - Industrial sectors - Population - GDP - Energy indicators --- ## Technologies Python GeoPandas Pandas NumPy Scikit-Learn XGBoost LightGBM CatBoost PySAL Folium Plotly Matplotlib Power BI --- ## Workflow Raw Data ↓ Cleaning ↓ Feature Engineering ↓ Spatial Analysis ↓ Spatial Autocorrelation ↓ Hotspot Detection ↓ Machine Learning ↓ Prediction Maps ↓ Interactive Dashboard --- ## Spatial Analysis The project performs: ✔ Global Moran's I ✔ Local Moran (LISA) ✔ Getis-Ord Gi* ✔ Spatial Weights Matrix ✔ Spatial Clustering --- ## Machine Learning Models Three algorithms were evaluated. - XGBoost - LightGBM - CatBoost Performance metrics - RMSE - MAE - R² --- ## Results The project successfully identifies: - Major African emission hotspots - Spatial clusters - Regional disparities - Future emission trends --- ## Repository Structure ```text data/ notebooks/ src/ figures/ maps/ models/ outputs/ ``` --- ## Dashboard The interactive dashboard allows users to - Explore emissions by country - Visualize hotspots - Compare regions - Analyse temporal evolut …