Adding all sample datasets to facilitate testing

[akorkovelos]

@ariane-millot let's try to document and add on the repo all the datasets used to run the different scripts. In the latest pull, I am missing several of the raster datasets that you probably used, thus I cannot run all elements of the code. We can use this issue to record all vector and raster datasets we use (name, source, justification) and then add them to a common folder preferably on GitHub. If the size surpasses GitHub's allowance (~50MB) then we:

• either limit the test areas from Cooperbelt to sth smaller or
• store everything on an external folder (e.g., G.drive, SharePoint) and add the link to the repo.

The latter would require a slight restructure of the scripts too. This will however allow easier testing moving forward, especially since we want to start merging the different pieces.

[ariane-millot]

I've set up this folder where we can put the data we use https://drive.google.com/drive/folders/1W6nQFr90Xc4AgIrZwL48Bg_9cftZI3c3?usp=sharing
Let's wait a bit before working on the list of the final datasets used! It would be better to have the methodology for buildings finalised before.