Feature Engineering Playbook

Encoding, scaling, outlier handling, feature creation and selection — branched separately for tabular, text, time-series and image data instead of averaged into one rule of thumb.

It covers categorical encoding, scaling, outlier treatment, missing values, feature creation and selection, branched separately for tabular, text, time-series and image data. It is written for practitioners whose model is underperforming for reasons that are probably not the model.

Every branch states the trade-off that decided it, so the recommendation you end on comes with the reasoning attached — something you can paste into a design note or defend in a review.