Which Way Does It Push?

A gym wants to know what makes members cancel. Each row describes one member, and each label is 1 if they cancelled and 0 if they stayed.

Write push_directions(rows, cancelled, names).

• names holds the column names, in the same order as each row.
• Fit make_pipeline(StandardScaler(), LogisticRegression(max_iter=1000)).
• Read the coefficients from the logistic regression step of the pipeline (named_steps["logisticregression"].coef_). There is one row of coefficients, with one number per column.
• A positive coefficient pushes towards class 1 (cancelling). A negative one pushes towards class 0 (staying).

Return a dict that maps each column name to "cancel" for a positive coefficient or "stay" for a negative one. No coefficient is zero in the tests.