How do you plan to compute that based on studies? If you could implement a nice way to do that based on actual study results it would be super interesting. As in not a bunch of hard-corded rules. Perhaps it could somehow be crowd-sourced. Not sure if there'll be enough people for that tho.
Also there should be a way to input what you actually it so it can self-adjust based on what you're actually doing rather than what you'd ideally be doing. As in "okay we can see that you don't really eat X so let's try Y".
But I guess then you'd loose 95% of potential users cause that's a pain to do.
There was this mental health company that would suggest the right medication based on statistics. Supposedly it was much more accurate than regular diagnoses that you'd get from a doctor. I can't find it thought ;C
