autogluon.features

What it is and what it does

autogluon.features is the feature engineering component of AutoGluon, AWS's automated machine learning framework. It handles the preprocessing and feature extraction steps that typically require manual tuning in ML pipelines, working with tabular, time series, and multimodal data. The package sits on top of numpy, pandas, scikit-learn, and autogluon.common, providing automated transformations and selection to prepare raw data for model training.

The package is designed to reduce boilerplate in the data preparation phase, fitting into AutoGluon's broader "3 lines of code" philosophy for end-to-end ML. It supports Python 3.10–3.13 across Linux, macOS, and Windows, and is actively maintained with recent commits and a large community (10596 repository stars). No known security vulnerabilities are recorded.

Use it for:

• Automatically engineer features from raw tabular datasets before training predictive models.
• Preprocess time series data for forecasting tasks within an AutoGluon pipeline.
• Extract and transform features from multimodal data (text, images, structured fields) for unified ML workflows.
• Reduce manual feature engineering effort in rapid prototyping and proof-of-concept projects.
• Integrate feature preparation into end-to-end AutoML workflows that also handle model selection and hyperparameter tuning.

Worth the install?

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

Automates feature engineering and preprocessing for machine learning pipelines, integrating with AutoGluon's broader ML automation framework to handle tabular, time series, and multimodal data preparation.

Yes, if you are building on AutoGluon or need automated feature engineering for tabular, time series, or multimodal data. The package is actively maintained, has no known vulnerabilities, and installs with low friction. It is most valuable as part of the full AutoGluon ecosystem; standalone utility depends on your pipeline architecture.

Install

autogluon-features on PyPI

pip install autogluon-features

uv add autogluon-features

poetry add autogluon-features

Installing autogluon.features

Before you install

Low friction install with a pure Python wheel. Actively maintained as of 8 days ago with a large repository (10596 stars) and recent activity. Depends on numpy, pandas, scikit-learn, and autogluon.common—all standard ML stack components.

License in practice

Apache-2.0 permissive license allows commercial and private use with minimal restrictions, making it suitable for both open and proprietary projects.

Quickstart

pip install autogluon.features

from autogluon.features import FeatureGenerator
from autogluon.common import space

# Prepare features for tabular data
generator = FeatureGenerator()
features = generator.fit_transform(data)

Requires Python 3.10–3.13; not compatible with older or newer Python versions.

Verify before relying

• Specific feature engineering algorithms and transformations supported beyond the runtime dependencies.
• Whether this package is typically used standalone or primarily as part of the full AutoGluon suite.
• Performance characteristics and scalability limits for large datasets.

Package facts

Evidence: autogluon_features-1.6.1-py3-none-any.whl

Tags

More Software Development packages

Further reading