Solar output predictor

The brief

Grid operators need to know tomorrow's solar generation to balance supply. Predict a plant's output from irradiation, temperature and time, compare regression models, and show where the forecast goes wrong. Good looks like: models compared with a simple physical baseline, honest time-based validation, and error analysis on cloudy days.

What you will build

1. Prepare weather and output data by time
2. Build a simple irradiation-based baseline
3. Train linear and gradient-boosted regression models
4. Validate on later days, not random rows
5. Analyse errors by weather condition

Free, open data to build it with

• Solar power generation data ↗ Kaggle

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.