Here are some maps
I made that show the Pronto stations overlaid on the Census block data for population, race, and income.
It isn't strictly true that they put the stations wherever density is highest. They covered a lot of high density areas, but they also covered some areas of lower density, while missing some denser areas. For example, they could have had less coverage in the lower density parts of Capitol Hill in favor of the super high density parts of Belltown. But perhaps they wanted to bridge to the U District. Though they could have bridged downtown to Belltown to Ballard and gotten some very high density that way.
But if you consider age, it probably makes sense to target the younger population in the U District than the older population in other high-density areas.
Racially, there are stations in some more diverse census blocks. What you can see they didn't do was extend the bike share far to the south into Rainier Valley. I suppose the only way they could have done so was to leave the U District out entirely and only put stations downtown and down into the CD and Rainier. Considering how low-density that area is, it would be a decision based wholly on race which ignored density entirely.
I don't know, maybe they should have. I tend to think that by targeting the most enthusiastic populations, on Capitol Hill and the U District, they helped to ensure the bike share is a success, making further expansion possible. Stretching out into a low density, diverse neighborhood could be seen as putting the health of the bike share at risk, and if it fails, everybody loses.
Anyway, if you play around with the maps, I think you'll see it's not such an easy problem. Erica Barnett said they failed
to align with major transit stations, but I think they hit most of them within a block or two. I'll add a tab to the maps this weekend showing where the bus and train stations are in relation to the Pronto stations.