What are the advantages this provides that specifically apply to working with visual end results?

There is one kind of iterative work I can see nodes/graphical coding have a great advantage: when you don’t know what you really want to do. You want to compose and reorder a bunch of operations or compositions thereof speculatively and just see what happens.

Moving code around and reorganizing it by trial and error is definitely something I can see visual programming excel at.