It combines animation of 3D models, including human bodies and machines, with perspective extrapolation and other effects. If someone can do that from photos, they also have the skills to make 3D animations from scratch. Lots of training, lots of work. There might be a fairly easy to use program somewhere that can try to do the perspective extrapolation, though. That could be interesting, though I am curious what distortion is involved, since there is no way to really see things that were not visible from the perspective of the original photo, so something tricky is being done – I think in order to get the illusion of a moving 3D perspective, they are sacrificing some of the actual data, one way or another. So if you really want to know what someplace looked like from a single photo, using one’s imagination while looking at the actual still photo is probably going to be more accurate, if less evocative.