Derek Delk: So are you saying keep the comment? The formulas are verifiably true without the comment or visualization, but the comment really helps connect each formula to portions of the figures in the visualization. However, Joerg Arndt says the comment is blog posting somehow.

Andrew Howroyd: The comment was hard to follow.
Yes polynomial identities abound, but if you can find 4 and everyone finds 4 then there will be so many. They don't strike me as useful and they all boil down to a(n) = (2*n + 1)^3 when expanded and then factored (which any CAS will do). Would someone who came to this sequence find these formulas useful? The answer is probably not. 

What I'd suggest is taking your png and turning it into a pdf with more information. (The trouble with 3d pictures is its quite hard to see what the shape is or how many spheres there are.). I am not sure how you can improve the illustration, perhaps some 2d slices (assuming there are stackable layers), but really it is up to you. (For me, arguing over what people put in their pdf's is mostly off limits). You can even include formula here that you think are relevant to the construction, but such formula don't necessarily have merit in the sequence entry itself.

Andrew Howroyd: A pdf will have the advantage of putting everything about this interpretation in one place instead of spreading between comments, a png and formulas. Also gives you more flexibility to present in a way that suits you. Using linked files is not uncommon. All papers are essentially this, but also see for example Fried's link in A391652.

Sean A. Irvine: This sequence has an incredibly simple definition. I don't think that any of this addition is helpful. Reverting.

Andrew Howroyd: If you give 4 formulas then nobody can know which one belongs with the visualization. At that point you might as well give 0.

Derek Delk: Never mind. It seems you only took issue with the comment. So I removed the comment, moved the formulas to the relevant section, and left the illustration as self-explanatory. Is this good?

Joerg Arndt: If we accepted this comment than this sequence would have potentially hundreds of similar ones, rendering it rather useless.
Also: the OEIS is not a blog.

Derek Delk: How is having various visual representations of a sequence problematic? I doubt accepting this would open the floodgates for hundreds more. The triangular numbers entry has only 7 different illustrations, while this entry has none so far. The fact that the figurate numbers for Escher's solid in bcc are odd cubes is surprising and interesting, and this insight yields new relations to other sequences.