Why should my discussion of the data be purely text-based (as I've had in my head). It should be multimodal. And this goes beyond a the use of enlarged type and so on as purely aesthetic visual effects. It should by critical. Here's an idea, then:
While this has been a useful bit of quick thinking-and-jotting-down, I suspect that I should probably focus on the data and see what types of visual approaches lend themselves to emerging themes, rather than thinking of visual approaches and then trying to find data to match.