Weather Forecasting Model

01 - The problem

Operational planning depended on forecasts that were too coarse for short-horizon scheduling.

02 - Approach

Cleaned and resampled multi-variable meteorological data, then trained and optimised an XGBoost forecasting model with iterative feature selection.

03 - Outcome

Improved predictive performance through optimisation, producing a forecast usable as a planning input rather than a curiosity.