lucaneviani/Geospatial-Analysis-of-Renewable-Energy-Power-Plants-using-Satellite-Data

Built a longitudinal panel dataset combining satellite imagery data used as proxy for economic activity, vegetation health and land cover transitions to study the local impact of solar and wind energy installations across multiple African countries. # Geospatial Analysis of Renewable Energy Power Plants in Africa **Satellite-Based Evaluation of Environmental and Economic Impacts of Solar and Wind Infrastructure** ## Overview Research project developed in collaboration with the University of Padua, investigates the local economic and environmental impact of utility-scale solar and wind power plants across Africa using satellite imagery and causal inference. Rather than relying on existing datasets, I designed and built an end-to-end geospatial data pipeline that collected, processed, and integrated satellite observations from multiple sources into a longitudinal georeferenced dataset. The final dataset was then used to estimate the causal effects of renewable energy infrastructure using state-of-the-art econometric methods. ## Academic Thesis This project was developed as part of my Master's thesis in **Economic Data Analytics**. The complete methodology, datasets, econometric framework, and empirical results are described in the full thesis document: 📄 **Read the full thesis** --- --- ## What I Built The core contribution of this project is the creation of a custom geospatial panel dataset from multiple heterogeneous data sources. The workflow included: - **Accessing satellite products** through NASA and ESA APIs - **Downloading multi-year Earth observation data** (2012–2024) - **Extracting satellite statistics** around thousands of renewable energy plants - **Cleaning and validating** raw geospatial datasets and geometries - **Performing spatial joins** and geometric buffer operations (500m to 20km) - **Building a longitudinal georeferenced panel dataset** with over 259,000 observations - **Running causal impact analysis** using modern Difference-in-Differences methods --- ## Data Engineering Pipeline To process satellite imagery at scale, I built an automated ETL workflow combining Python and Google Earth Engine: ```mermaid flowchart TD A["⚡ Renewable Power Plant Database (RePP Afric …