Fascinating. Also fascinating to see the relative importance they attribute to wildly different geographic abstractions – US/Canada/Australia called out, meanwhile Europe and Asia bucketed together, despite those being the more instructive studies for how to operate mass transit given how well-developed it is in some of their major cities. I'd love to have seen more granular data for those two continents as I suspect it will reveal interesting trends and reasons for fare capture that this analysis doesn't identify.

> His idea is to turn hub-and-spoke transit, in which most lines go downtown, into a grid so that people can get anywhere in the region with just one change of bus

Anyone who has played any simulation of building mass transit systems (e.g. see the game "Mini Metro") knows it's far more efficient if the system is designed such that any origin-destination pair can be reached with only a single change.