What actually drives a solar panel’s yield? This project starts with a uni- and bi-variate analysis, complemented by a PCA (principal component analysis), to identify the most determining features — notably the number of hot spots and the panel’s average temperature.
From there, several regression models are compared to predict the yield. The best result comes from a Support Vector Regression (SVR):
Tools
- Pandas / NumPy — data preparation
- Seaborn / Plotly — visualization and exploratory analysis
- Scikit-learn — modeling and comparison (including SVR)