skrub

What it is and what it does

skrub is a Python library that bridges the gap between raw dataframes and machine learning models by providing transformers and utilities for common data preparation tasks. It sits in the scikit-learn ecosystem and works with pandas DataFrames and numpy arrays, offering tools for feature engineering, encoding, and data cleaning that are typically needed before training models.

The library depends on numpy, pandas, scikit-learn, scipy, and visualization tools (matplotlib, pydot) for its operations. It is actively maintained, supports Python 3.10 through 3.14, and has been in production use since late 2023. The package is designed to integrate with sklearn's Pipeline API and other standard ML workflows, making it a natural fit for teams already using those tools.

Use it for:

• Prepare messy tabular data with missing values and mixed data types for supervised learning.
• Encode categorical features and handle string columns in a sklearn-compatible way.
• Join multiple dataframes and align them for feature engineering in a machine learning pipeline.
• Transform raw CSV or database exports into clean feature matrices ready for model training.
• Build reproducible data preprocessing workflows that integrate with sklearn Pipelines.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

skrub prepares and transforms dataframes for machine learning by handling common data cleaning and feature engineering tasks that scikit-learn and pandas workflows typically require.

Yes. skrub is actively maintained, has no known vulnerabilities, low install friction, and fills a genuine gap in the sklearn ecosystem for dataframe-native preprocessing. It is well-suited for teams doing tabular machine learning with pandas and scikit-learn. Install it if you regularly work with raw dataframes and want to avoid writing custom preprocessing code.

Install

skrub on PyPI

pip install skrub

uv add skrub

poetry add skrub

Installing skrub

Before you install

Low friction: pure Python wheel with well-established dependencies (numpy, pandas, scikit-learn, scipy). Active maintenance with a recent release 39 days ago and steady repository activity.

License in practice

BSD-3-Clause is permissive; you can use, modify, and distribute skrub freely in commercial and private projects with minimal restrictions.

Quickstart

pip install skrub

import skrub
from skrub import Joiner
# Use skrub transformers in a sklearn Pipeline or standalone

Requires Python 3.10 or later; scikit-learn, pandas, and numpy must be installed.

Verify before relying

• Specific transformers and their capabilities (e.g., handling missing values, encoding strategies, joining logic) are not detailed in the fact sheet.
• Performance characteristics on large datasets or typical dataframe sizes are not documented here.
• Whether skrub integrates directly into sklearn Pipelines or requires wrapper code is not explicit in the fact sheet.

Package facts

Evidence: skrub-0.10.0-py3-none-any.whl

Tags

More Libraries packages