TIMEX

TIMEX (referred in code as timexseries) is a framework for time-series-forecasting-as-a-service.

Its main goal is to provide a simple and generic tool to build websites and, more in general, platforms, able to provide the forecasting of time-series in the "as-a-service" manner.

This means that users should interact with the service as less as possible.

An example of the capabilities of TIMEX can be found at covid-timex.it
That website is built using the Dash, on which the visualization part of TIMEX is built. A deep explanation is available in the dedicated repository.

Installation

The main two dependencies of TIMEX are Facebook Prophet and PyTorch. If you prefer, you can install them beforehand, maybe because you want to choose the CUDA/CPU version of Torch.

However, installation is as simple as running:

pip install timexseries

Get started

Please, refer to the Examples folder. You will find some Jupyter Notebook which illustrate the main characteristics of TIMEX. A Notebook explaining the covid-timex.it website is present, along with the source code of the site, here.

Documentation

The full documentation is available at here.

Contacts

If you have questions, suggestions or problems, feel free to open an Issue. You can contact us at:

• alessandro.falcetta@polimi.it
• manuel.roveri@polimi.it