> The Bay Area has a population density many times larger than a long list of places with vastly better public transit.
Well of course the distribution matters greatly too (indeed, maybe it's all that matters).
If the overall population density is the same, but the population is mainly clustered around specific nodes (towns, neighborhoods) and routes (streets), providing good transit will be vastly easier than if the population is evenly distributed as seems to be the goal of 1950s American suburbia.
The clustering of population, of course strongly follows the structure of the dominant transportation method, so such effects also tend to be self-perpetuating.
If anything, I think the Bay Area would be substantially easier to cover well than a lot of places I've lived or travelled.
Sure, the US car centric model isn't ideal for public transport, but the Bay Area does have substantial high density clusters by the standards of e.g. Norway where I'm from.
It probably wouldn't have been if your population density had been comparable. But the Oslo metro region for example, outside of a tiny city core, consists largely of tiny little towns with mostly single or dual residency houses, terraces and a few low apartment buildings. Larger apartment blocks are found some places too, but most commonly in the centre of Oslo. There are large stretches of pretty much empty space - forest or farms. Density for the region as a whole is driven up substantially by the city of Oslo itself.
The Bay Area on the other hand, at least has most of the population clustered around a handful of major transport links, even though the towns spread out more than a similarly sized city usually would in Europe. Increasing frequency of Caltrain, running it later, and adding a handful of extensions to Caltrain and Bart, coupled with a properly unified bus network, would make all the difference, I think.
Well of course the distribution matters greatly too (indeed, maybe it's all that matters).
If the overall population density is the same, but the population is mainly clustered around specific nodes (towns, neighborhoods) and routes (streets), providing good transit will be vastly easier than if the population is evenly distributed as seems to be the goal of 1950s American suburbia.
The clustering of population, of course strongly follows the structure of the dominant transportation method, so such effects also tend to be self-perpetuating.