woodwork

What it is and what it does

Woodwork is a data typing library that extends pandas DataFrames with a semantic layer for machine learning. It lets you assign logical types (Integer, Categorical, DateTime, NaturalLanguage, etc.) and semantic tags to columns, then query and filter your data based on those annotations. The library automatically infers types from underlying data when you don't specify them, and it stores metadata alongside your DataFrame for use in downstream ML workflows.

The package is designed as a common typing namespace for Featuretools and EvalML, so if you're using those tools, Woodwork provides a standardized way to communicate data structure and meaning. It depends on pandas, scikit-learn, scipy, numpy, python-dateutil, and importlib-resources—all standard data-science libraries—so installation is straightforward. The project is maintained by Alteryx but has not seen active development since mid-2024.

Use it for:

• Annotate DataFrame columns with logical types and semantic tags before passing data to Featuretools for automated feature engineering.
• Store and communicate column metadata (e.g., which columns are numeric, categorical, or contain personal names) for reproducible ML pipelines.
• Filter and select DataFrame subsets by logical type or semantic tag without manual column name lists.
• Standardize data typing across multiple DataFrames in a machine learning project to ensure consistent downstream processing.
• Integrate with EvalML for automated machine learning workflows that rely on consistent column type information.

Worth the install?

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

Woodwork adds a typing layer to pandas DataFrames, letting you annotate columns with logical types and semantic tags for use in machine learning workflows and data science tools.

Yes, if you are using Featuretools or EvalML and need a standardized way to annotate and manage DataFrame column types. The low install friction and permissive license make it a straightforward addition. However, be aware that the project is aging—last release was May 2024 and there have been no commits since September 2025—so verify compatibility with your current versions of pandas and scikit-learn before relying on it for new projects.

Install

woodwork on PyPI

pip install woodwork

uv add woodwork

poetry add woodwork

Installing woodwork

Before you install

Low install friction with 6 common runtime dependencies (pandas, scikit-learn, scipy, numpy, python-dateutil, importlib-resources). Maintenance status is aging—last release was in May 2024 and no commits since September 2025, suggesting the project is stable but not actively developed.

License in practice

BSD 3-Clause License is permissive; you can use, modify, and distribute Woodwork freely in commercial and private projects as long as you retain the license notice and disclaimer.

Quickstart

import pandas as pd
import woodwork as ww

df = pd.read_csv('data.csv')
df.ww.init(name='my_data')
df.ww.set_types(logical_types={'col_name': 'Integer', 'category_col': 'Categorical'})
filtered = df.ww.select(include=['numeric'])

Verify before relying

• Whether the aging maintenance status (last release May 2024, no commits since Sept 2025) affects compatibility with recent pandas or scikit-learn versions.
• Whether Woodwork's integration with Featuretools and EvalML remains current given the maintenance gap.

Package facts

Evidence: woodwork-0.31.0-py3-none-any.whl

Keywords: data science, machine learning, typing

Tags

More Software Development packages