Classification with scikit-learn: telling two populations apart

Afterwards you can fit a scikit-learn classifier, score it on a held-out test set, tune it by cross-validation, and chain a scaler and model in one pipeline.