Because mathematics is completely divorced from the empirical it is not constrained by the need for evidence.
A sentence which is absolutely false. Mathematics is the tool of the empirical mind, and it started when someone started counting beans. Or potatoes. The details are somewhat unclear

. Mathematics depends upon the empirical.
While mathematical axioms must be stated, altering the axioms simply leads to new mathematical approaches. No mathematical system will ever be complete, but mathematical constructs are objectively real in the way the charge on the electron is real - they are not constructs the way color is.
Exactly, this I can agree with. Mathematical constructs are as "real" as the charge of the electron. That is a nice proposition.
Anyone who exists in the material world can derive the concept of a sphere. They don't need to find the Sphere Species and dissect their brains and model them
This is an argument that states that the definition of sphere is much more simple than the definition of color. We can agree that this quantitative difference will translate to a qualitative one, because that distinction is useful for us humans with our little brains.
I agree, it absolutely can, but there is no way that an alien species observing its environment would arrive at the human conception of color without
1) evolving the same visual system in parallel
2) deciding to **** around and simulate a bunch of visual systems
Number 2 is clearly what is being discussed here, and it is true. You're arguing that spheres are "different" because they are easier. But mathematics doesn't stop being mathematics just because it is more complex.
But I wasn't even discussing "spheres" but "shapes". A shape is a perception of a figure that you see with your eyes. This perception is tainted by many things, for example how your eyes work the things that are around you into tridimensional shapes that you can platonically caricaturize, like say, a cube. An alien that could "see" the space between and within atoms, between the core and the electrons, etc., could well develop a completely different geometry. A sphere could be a really strange (but definable) geometry for it. Or it could be not.