OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Data & Analytics

Skforecast

Skforecast is an open-source Python library for building time-series forecasting workflows with scikit-learn-compatible estimators, statistical methods, Keras-based deep learning and foundation models. It provides tools for forecasting one or many series, incorporating exogenous predictors, and backtesting model approaches.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Skforecast — OpenViST project profile in AI Data & Analytics.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Skforecast when a Python forecasting project needs a common workflow for fitting, predicting and backtesting across compatible machine-learning estimators and other supported model families.

Who it is for

  • Python practitioners building forecasts from historical time-series data.
  • Teams comparing machine-learning, statistical and foundation-model forecasting approaches.
  • Users working with multiple series or external predictors such as temperature and holidays.

Documented uses

  • Fit a forecasting model to a time series and generate future predictions.
  • Compare forecasting approaches through backtesting.
  • Forecast multiple series with recursive or direct strategies.
  • Include exogenous predictors in a forecasting workflow.
  • Use a supported foundation model for zero-shot forecasting.
Before you choose

Requirements & limits

The repository documents installation through pip and conda-forge. It lists Python 3.10–3.14 and a BSD-3-Clause license.

  • The library is Python-based; its README lists Python versions 3.10–3.14.
  • Model families can require additional packages: the Chronos-2 example calls for chronos-forecasting, while optional dependencies for statistical models, plotting and deep learning are described separately.
  • Backtesting supports different evaluation configurations, including refits, gaps, fold strides, and fixed or expanding windows.
  • The same forecasting library covers compatible regressors, statistical methods, Keras-based deep learning and foundation-model workflows.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Apply a foundation model to the demo series

Use the README's zero-shot Chronos-2 workflow to request a 12-step forecast without training the foundation model.

Before you start

  • Python environment with Skforecast installed
  • The chronos-forecasting package, installed separately
  • A demo series loaded through Skforecast

Steps

  1. Install the additional package with the command below.

    Documented command
    pip install chronos-forecasting

  2. Import FoundationModel, ForecasterFoundation and load_demo_dataset from the documented modules.

  3. Load the demo series and create a ForecasterFoundation using FoundationModel("autogluon/chronos-2-small").

  4. Pass the series to fit to store its context, then call predict(steps=12).

Expected outcome

A 12-step forecast from the Chronos-2 example, with no model training in the documented workflow.

Guide limitations

  • The recipe documents this specific Chronos-2 example; it does not establish requirements for every foundation model.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
skforecast/skforecast · Official homepage title · retained in its source language
How we review sourcesSuggest a correction