We don't understand cognition, consciousness, and much else. What we assume about why the data is generated as it is critically affects how we interpret that data, and simply crunching numbers is nothing more than Sudoku.
Which is why I avoid assuming. I simply say I don't know. I don't know if there really are individuals. The science (physics) says there is no free will for example. I also don't know whether or not consciousness is an illusion as some philosophers say or whether it's quantum as some mathematicians say.
All I can see are behaviors. So the data we have does not include what they call "private events". We cannot know anything other than behaviors and from this we have data. Social physics at best can detect behaviors from the data or in some cases maybe predict future behavior statistically but it cannot work magic. It could be because we simply don't have enough data or it could be that social physics just isn't as powerful as some of the mathematicians seem to think.
I'm not going to defend or attack social physics but if it proves to be useful then I'll make use of it out of pragmatism.