Agreed, this looks like a classic example of "correlation != causation".
It seems like this argument taken to it's final conclusion would be that startups gravitate towards cities with better transit than those that don't, which isn't the case given the Austin and San Jose examples.