Naive benchmark forecast methods using Stan as a backend

Project: tablespoon

In the Python community, a lot of effort is put into sophisticated forecasting methods, but naive probabilistic forecasting has been overlooked.

tablespoon or tbsp provides easy-to-use methods to fill that gap. It is expected that this package may be used as a compliment to sophisticated methods. This package is exceptionally ordinary!

This package uses cmdstanpy as the backend to take advantage of the Stan sampler and because the models get compiled. It ends up running really fast!

Though this is in beta, I have started using it at work and the results have been fantastic!

I ❤️ Stan!

Contributions Welcome!


  author={Alex Hallam},
  title={{tablespoon}: {Time-series Benchmark methods that are Simple and Probabilistic},
  note={Version 0.1.8,