EuroSciPy 2024

sktime is the most widely used scikit-learn compatible framework library for learning with time series. sktime is maintained by a neutral non-profit under permissive license, easily extensible by anyone, and interoperable with the python data science stack.

This tutorial gives a hands-on introduction to sktime, for common time series learning tasks such as forecasting, starting with a general overview of the package and forecasting interfaces for uni- and multivariate forecasts with endo-/exogeneous data, probabilistic forecasts, and forecasting in the presence of hierarchical data.

The tutorial then proceeds to showcase some of the newest features in 2024, based on a hierarchical demand forecasting use case example: support for foundation models, hugging face connectors, advanced support for hierarchical and global forecasts, and integration features for creating API compatible algorithms and sharing them via the sktime discoverability tools.

Foundation models are here for forecasting! This will conclusively solve all forecasting problems with a one-model-fits-all approach! Or … maybe not?

Fact is, an increasingly growing number of foundation models for time series and forecasting hitting the market.

To innocent end users, this situation raises various challenges and questions. How do I integrate the models as candidates into existing forecasting workflows? Are the models performant? How do they compare to more classical choices? Which one to pick? How to know whether to “upgrade”?

At sktime, we have tried so you don’t have to! Although you will probably be forced to anyway, but even then, it’s worth sharing experiences.

Our key challenges and findings are presented in this talk – for instance, the unexpected fragmentation of the ecosystem, difficulties in evaluating the models fairly, and more.

(sktime is an openly governed community with neutral point of view. You may be surprised to hear that this talk will not try to sell you a foundation model)