StatsForecast is an open-source Python library that provides a wide range of statistical and econometric forecasting models. It is one of the most widely used forecasting libraries in the field, with over 20 million downloads and more than 4,000 GitHub stars.

In this workshop, we’ll take a look under the hood of StatsForecast. We’ll explain its architecture and what makes it so fast and scalable. We will also examine the implementations of the main model families: ARIMA, ETS, Theta, MSTL, baseline models, and models for sparse and intermittent series.

StatsForecast was recently featured in the Python edition of Forecasting: Principles and Practice, a key reference in the field. This open-source book is available at and represents an important part of StatsForecast’s history, as the library began as a Python implementation of many models from the forecast R package.

We’ll conclude with a look ahead at planned features and upcoming models for StatsForecast.