The use of ThingLInk is an interesting strategy, but I think the current composition of images relies too heavily on the accompanying words to convey anything you are trying to convey. I think that might be because you are thinking too literally about remediating your text object into image. If you consider what the major questions are the essay you've composed (Who is the man? How do we stick it to him? Why do we stick it to him?) could you break this up into multiple pieces that examine each of these questions? Does every image need to have a popup? (other than for attribution purposes) For revision, consider a move away from a literal transposition of the essay to a collection of images that gets at the major questions or argument of the essay.