Sanity check graph takes a long time to load

[matt-graham]

Submitting as part of JOSS review openjournals/joss-reviews/issues/3651

I really like the sanity check example in the documentation, however I think it may be worth using a different location than London to source the example graph as running

graph = ox.graph_from_place('London, UK')

takes around 8 minutes on my laptop and the lack of any progress updates from the osmnx package when executing graph_from_place made it a bit unclear initially if the interpreter had crashed. Swapping for a location that produces a smaller graph would still perform the same function of illustrating the inference algorithm is computing something sensible but without requiring so much patience from the user! If you prefer to keep the current example, I think it would then be worth at least putting a warning to expect the osmnx functions to take a long time to run.

[SamDuffield]

As a large portion of the anticipated use cases is in urban environments (most likely major cities such as London) I think leaving the example as is but adding a precautionary comment about osmnx loading times is the most useful for the user. (Also the sample_route function works better for larger areas as it is less likely to terminate early on reaching the edge of the area.)

I have added said warnings to the docs!

[matt-graham]

As a large portion of the anticipated use cases is in urban environments (most likely major cities such as London) I think leaving the example as is but adding a precautionary comment about osmnx loading times is the most useful for the user. (Also the sample_route function works better for larger areas as it is less likely to terminate early on reaching the edge of the area.)

Those both seem good rationales for keeping the current example, thanks for the clarifications!

I have added said warnings to the docs!

👍