skrub

skrub is a Python library designed to simplify machine learning workflows when working with tabular data (dataframes). It provides tools and utilities to preprocess, transform, and integrate dataframe-based data into scikit-learn pipelines, making it easier to handle real-world datasets with mixed types, missing values, or categorical features. The library includes specialized encoders, transformers, and helpers for seamless compatibility with pandas dataframes and scikit-learn estimators. It supports installation via pip or conda and offers extensive documentation, examples, and community resources (e.g., Discord, GitHub Discussions).