Check and document the generalisability of the workflow

[cforgaci]

This is an issue raised as part of ropensci/software-review#718

As pointed out in the review,

The workflow is presented in a way that should generalise to other cities or rivers, provided there is sufficient OSM coverage and DEM availability. [...] However, I encountered some difficulties when attempting to reproduce the workflow in a randomly selected Italian city. It was not as smooth as I expected.

We should properly document any limitations one might encounter when attempting to delineate with rcrisp on a random city.

Also, while following up on the rest of the review, we should keep in mind any hindrances to or improvements on generalisability.

[cforgaci]

This could be implictly addressed by estracting data retrieval as suggested in point 2 of #386

[cforgaci]

Addressed in #370