Preparation steps and a model packaged as one object, so new data is treated exactly the same way as training data. Covered properly in Session 2.

Examplemake_pipeline(StandardScaler(), LogisticRegression()) scales and models in one fit.