On Vimeo he says that everything on the computer screen is animated without any replacements. So I'm guessing that yes, it's all just street view.
You can clearly see that it's fluid video, not the smudgey, morphing transition used in street view.
Not sure if that is what they actually did, but playing around with it a bit on my own it looks like you could get some decently fluid looking motion by taking screenshots of each "frame" that street view provides.
According to a quick calculation that would give you a minimum apparent speed of 170 mph. (15 f/s * 5 m/f * 3600 s/h / 1600 mph/m/s)