Back to case studies
ML · Data Science

XGE

End-to-end XGBoost pipeline predicting vehicle CO₂ emissions with feature engineering, gain-based explainability, and an interactive dashboard.

XGE 1XGE 2

Problem

Regulators and OEMs need a transparent way to estimate emissions for a given vehicle spec without trusting a black box.

Approach

  1. Cleaned the vehicle emissions dataset; label-encoded categoricals and scaled 8 features with RobustScaler.
  2. Tuned XGBoost with a 256-combination GridSearchCV sweep — 1,280 five-fold fits.
  3. Surfaced gain-based feature importances and residual diagnostics so every prediction is auditable.

Results

R²
0.95
MAE
9.0 g CO₂/km
RMSE
13.0 g CO₂/km