Geolocation features in demand forecasting

[polkadot21]

I am really thankful for the package and nice tutorials. Great job, team!

Working on a similar task as in the Stallion example (https://pytorch-forecasting.readthedocs.io/en/stable/tutorials/stallion.html), I am thinking about using geolocation of the stores as a feature (two-dimensional vector (latitude, longitude) for every location. My assumption is that the demand, discounts, etc in the next store (or stores) may affect the demand in the store of interest.

Are there any good practices to leverage geolocation information in TFT? Any help would be highly appreciated.