Build a baseline classification model from messy data

You are working inside an interactive notebook environment, such as Jupyter, and receive a raw tabular dataset for a binary classification problem. The target column y only takes the values $$0$$ or $$1$$. Properties of the data: Columns are a mix of numeric and categorical features. Some cells are missing; categorical strings may contain inconsistent markers such as "NA" or empty text; numeric columns may contain outliers. Some columns are identifiers, for example user_id…