Basic Feature Engineering

Contents

Missing Data

1. EDA: Using regular expressions, visualization techniques, and the LittleMCAR function from R to explore missingness and determine its mechanism
2. DataWig: a deep learning library for the imputation of missing categorical data

Categorical Data

1. Cardinality: Quantify and reduce cardinality
2. Encoding: Popular and evidence-based encoding techniques and potential pitfalls. Methods for both low and high cardinality features

Feature Scaling