Renewable Power Forecasting

Inspiration

Grid operators need short-horizon forecasts of solar and wind output to balance supply. This project trains and evaluates both forecasters on seasonal generation data.

What it does

Predicts photovoltaic and wind power generation from weather features; evaluation plots (actual vs predicted, gallery) come straight from the trained models.

How I built it

• Python, scikit-learn regressors (see ML train.py, preprocessing in preprocecssing.py)
• Seasonal CSVs for solar (spring/summer/autumn/winter) and wind
• Saved models: solar_power_model.pkl, wind_power_model.pkl + scalers; test results CSVs in repo

Results (re-ran locally)

• Solar: MAE 97.99, RMSE 216.81, R2 0.9736
• Wind: MAE 57.26, RMSE 135.03, R2 0.9874

Challenges

Seasonal distribution shift between train and test splits.

What I learned

Feature scaling + simple regressors go far on clean energy data.

Links

• Code: https://github.com/Priyanshu85277/renewable-energy

Built With

• energy
• forecasting
• machine-learning
• python
• scikit-learn