All these models ignore how insects actually zero in on a smell. For a fly outside, the source of the smell is the gap in the window. What looks like random direction changes is actually a search pattern for that source.
You are right. Even though it is probably a great exercise to improve your skills as data scientists, I'm afraid that without understanding how the insects actually "work/think", it is nothing more than garbage in, garbage out (GIGO).
There is presumably also a falling-off of the signal with distance. Flies may have a good sense of smell and all, but it's not infinitely good. Which is too bad, because if they did, then all the flies in NYC (or I suppose maybe the world) would all be in one big clump near the best (to a fly) smelling object. Although I guess if you were that object, then it would be not as good.
> All these models ignore how insects actually zero in on a smell.
It seems to me that the author is not ignoring that:
> While the flies now obey the laws of aviation, there's still something that seems missing. Normally, flies want to get indoors. They need to get indoors. So to simulate this, I added an attractor (maybe a rotting piece of fruit) at a position A indoors to lure the flies.
The model where the fly knows where the source of the smell is located the weird outcome of the flies targeting the pane in the middle instead of the sides where the smell escapes.
Same principle behind a DIY fruit fly trap (glass of juice with a plastic cover with a few holes poked in it).