Nixtla/statsforecast

Lightning ⚡️ fast forecasting with statistical and econometric models.

What it solves

StatsForecast provides a high-performance library for univariate time series forecasting. It addresses the problem of existing Python alternatives for statistical models being too slow, inaccurate, or unable to scale to millions of time series in production environments.

How it works

The library implements a wide array of statistical and econometric models—such as AutoARIMA, AutoETS, AutoCES, and Theta—optimized for speed and accuracy. It uses a familiar sklearn-style .fit and .predict syntax to make it easy to integrate into existing workflows. To handle massive datasets, it offers out-of-the-box compatibility with distributed computing frameworks like Spark, Dask, and Ray.

Who it’s for

Data scientists and ML engineers who need to generate forecasts for large collections of time series data and require high performance, probabilistic forecasting, or the ability to scale to millions of series.

Highlights

  • Extreme Speed: Up to 500x faster than Prophet and 20x faster than pmdarima.
  • Scalability: Capable of fitting 10 benchmark models on 1 million series in under 5 minutes using Ray.
  • Comprehensive Model Suite: Includes automatic forecasting tools, ARIMA family, Theta family, MSTL for multiple seasonalities, GARCH/ARCH for volatility, and specialized models for sparse or intermittent demand.
  • Advanced Features: Supports exogenous variables, static covariates, prediction intervals, and anomaly detection.

Related

  • Project
    skforecast/skforecast

    A Python library for time series forecasting that integrates scikit-learn compatible models, statistical methods, and foundation models into a unified forecasting workflow.

  • Project
    mikekeith52/scalecast

    A time series forecasting library that provides a uniform interface for ML and deep learning models, simplifying model selection, validation, and pipeline creation.

  • Project
    Nixtla/neuralforecast

    A library of over 30 state-of-the-art neural forecasting models providing a user-friendly, scikit-learn style API for high-performance time-series prediction.

  • Project
    Nixtla/mlforecast

    A scalable framework for time series forecasting that allows any scikit-learn compatible machine learning model to be used for forecasting millions of series.

  • Project
    unit8co/darts

    A Python library for user-friendly time series forecasting and anomaly detection, offering a unified API for models ranging from ARIMA to deep neural networks.