alteryx/featuretools

An open source python library for automated feature engineering

What it solves

Featuretools is designed to automate the feature engineering process, which is often a manual and time-consuming part of machine learning workflows. It allows users to automatically create a single table of features from multiple related tables of data, transforming raw data into a feature vector for each target entity (e.g., customers) that can be used for machine learning models.

How it works

The library uses a process called Deep Feature Synthesis (DFS), which automatically calculates features across multiple tables. It utilizes "primitives"—the building blocks for creating features—to perform calculations. These primitives can be the built-in ones provided by the library or custom-defined by the user. It can handle multi-table datasets, such as timestamped customer transactions, and can scale to large datasets using Dask support.

Who it’s for

Data scientists and machine learning engineers who need to efficiently generate features from relational data for their ML pipelines.

Highlights

  • Automated Feature Engineering: Uses Deep Feature Synthesis to automatically generate feature matrices from multi-table datasets.
  • Custoimzable Primitives: Includes a variety of built-in primitives and optional add-ons for NLP and premium features, while allowing users to define their own custom primitives.
  • Scalability: Supports Dask for running DFS on large datasets with multiple jobs.
  • Multi-table Support: Specifically designed to handle complex relational data structures (entitysets).

Related

  • Project
    feature-engine/feature_engine

    A Python library providing Scikit-learn compatible transformers for comprehensive feature engineering and selection to improve machine learning model performance.

  • Project
    MTSWebServices/RecTools

    A Python library that streamlines the creation of recommender systems by providing a unified interface for various models, from matrix factorization to advanced Transformers.

  • Project
    winedarksea/AutoTS

    An AutoML package for Python that uses genetic algorithms to automatically find the best models and preprocessing for high-accuracy, large-scale time series forecasting.

  • Project
    sdv-dev/SDV

    A Python library for creating tabular synthetic data using machine learning to emulate real-world patterns while maintaining privacy.

  • Project
    materialyzeai/maml

    A Python package that provides high-level interfaces for materials science machine learning, enabling the conversion of materials data into features for property prediction and energy surface modeling.