Fi-g/sa-solar-farm-location-predictor

A data science project identifying optimal solar farm locations for underserved rural communities across South Africa # South African Solar Farm Location Predictor A data science portfolio project identifying optimal solar farm locations for underserved rural communities across South Africa. ## Project Overview This project combines village-level settlement data, national grid performance data, and weather station observations to: - Score 76941 South African settlements for solar farm suitability - Forecast national electricity demand through to 2030 - Identify 8 priority solar farm locations across 3 provinces - Calculate the minimum number of farms needed to reach 25% renewable energy penetration by 2030 ## Dashboard View the interactive dashboard here: SA Solar Farm Location Predictor Dashboard Or view the static PDF export: Dashboard PDF ## Key Findings - Thabazimbi (Limpopo) ranks as the top priority location, serving 108202 people within its 50km transmission radius - The Northern Cape dominates suitability ranking due to high solar irradiance, low cloud cover, and low security risk - A minimum of 221 solar farms are needed by 2030 to reach the 25% RE penetration target - South Africa's grid demand has declined consistently since 2021 due to accelerating rooftop solar adoption ## Project Structure | Folder | Description | |---|---| | notebooks / | Jupyter notebooks for each project phase | | dashboard / | Power BI dashboard (.pbix) | | report/ | Full portfolio report | | outputs/ | Visualisation files (PNG and HTML) | | data/ | Raw and processed data (not included - see Data Sources) | ## Data Sources | Dataset | Source | Access | |---|---|---| | National grid data | Eskom | Formal data request | | Weather station data | South African Weather Service (SAWS) | Formal data request | | Settlement data | DRE Atlas | energydata.info | ## Tools and Technologies Python | SQL | SQLite | Power BI | Prophet | GeoPandas | Folium ## Notebooks | Notebook | Description | |---|---| | phase1_eskom_data_cleaning | E …