I cannot accept an analogy of the Universe as a 2D surface of a sphere and I think it‘s ludicrous woo to do so. The Universe is always spatially 3D. Nothing more, nothing less.
People like Albert Einstein, Hermann Minkowski, Hendrik Lorentz, Henri Poincaré, Carl Sagan, Stephen Hawking and Michio Kaku disagree with you so I don't think it can be so casually dismissed as "ludicrous woo".
Let me see if I can't lay it out in a bit more detail: First of all, current models for the structure of the universe suggest that there are a good bit more than three spatial dimensions at play. The problem is that it is practically impossible for us to visualize dimensions beyond the three (four, actually, including time) that we think in.
We can describe the fourth dimension rather easily. Define a point in space, then expand that point to a 10 cm line segment. That is one dimensional space consisting of length only. Even if it loops around on itself it is still a closed one dimensional space, but it is now without any internal boundary. It is finite but boundless within its single dimension. A one dimensional inhabitant could set off in one direction and return to its starting point from the opposite direction.
Now let's define a two dimensional space. Go back to our original point and draw a 10 cm line segment radiating from the original point at a 90 degree angle to the first segment. We have now defined a two dimensional space with length and width only. If we wrap it around on itself into a sphere we have the same situation as we did with the one dimensional ring. The space is two dimensional, yet it is boundless
and finite. "Mr. Square" could set out in one direction looking for the end of the universe and discover that after traveling a certain distance he will come back to his starting point from the opposite direction.
Now lets move on to three dimensions. The model is the same: we take our original point and draw another line segment radiating from it at 90 degrees to
both of our earlier lines. We squared one dimension and got two, then we squared two and got three. Now we have a space that is 10 cm in length, 10 cm in width and 10 cm in height. We have a tree dimensional cube.
But can we wrap the cube around upon itself in the same way as the line and the square? The answer is "yes", but we can't visualize it. Why not? Because we would have to be able to visualize a point with four lines radiating from it with each line at 90 to the other three. That is the fourth dimension put very simply. If we could fold the cube around in the fourth dimension, just as we folded the line into the second dimension to make a ring and folded the square into the third dimension to make a sphere, we would have a space in which a traveler could set off in any direction and return to the point of origin without ever turning or reversing. But again, we can't visualize it. If you've ever noticed Paulhoff's avatar then you have seen a representation of the three dimensional shadow of a four dimensional cube (a hypercube for short) rotating in space. Just as a three dimensional object casts a two dimensional shadow on a surface, a hypercube will cast a three dimensional shadow in three dimensional space. A real hypercube would look like Paulhoff's avatar, but all the line segments would be the same length and all the angles would be 90 degrees.
This inability to visualize higher dimensions (and yes, you can square the hypercube) is the reason that we have to resort to analogies based on two dimensions. This isn't even a new concept. One of the most famous examples being the 1884 book
Flatland: A Romance of Many Dimensions by Edwin Abbott Abbott (yes, Abbott squared). It contains a lot of contemporary social satire, but the basic model is still considered very valid. I suggest doing a bit of reading on the subject of flatland in general. You may learn some fascinating things. For example: a sealed cube is actually wide open to four dimensional space, just as a sealed square (at least, so seeming to a Flatlander) is open to three dimensional space.
This is far from a comprehensive explanation, but I think you can get the basic drift from it.