Metajunks experiment uses the perfect sphere/parallel light ray model that also uses the assumptions that the position of the equator is equi-distant to the poles and that latitudes are equidistant from the equator. This fits his agenda.
It doesn’t use or assume anything.
People in different latitudes took a measurement of the angle of the sun.
All he’s done is then map out what that would look like we’re the earth flat. Yes, he’s assumed that the degrees of latitude are equidistant, if that’s not correct in your FE model then you could try mapping the observations on to your model.
The point is the lines point all over the place, so where is the sun? The sun must physically be in a place so the lines should all point to a common place, but they do not.
So he’s tried to map the observations on to a spherical earth and with that model they all point in the same direction and are parallel which implies a distant sun.
Flat Earthers often misunderstand the oblateness thing. While the earth IS oblate it only bulges slightly at the equator. If you look at photos of the whole earth from space you can’t see it with the naked eye (although if you count pixels across and up and down on a good enough resolution picture you’ll see a difference). So the oblate nature of the earth isn’t going to change the way he’s mapped the results on a sphere enough to make a difference.
This is not "experimental proof” of a globe earth and the scientific method isn’t about proving anything. It’s about testing and disproving things. This experiment shows that a globe earth model fits the observations better than a flat earth one - or the flat earth one he has used, as I said maybe there is a flat earth model where the lines do point to a common point. Or maybe there’s another explanation, light bending perhaps.
If the sun is close then you should be able to take observations from a few different places and triangulate to calculate its distance. I have yet to see any evidence that experiment has been done.