[khafra]

In the book "Surely You're Joking, Mr. Feynman," he talks about needing a concrete example of any abstract system he was taking a new look at; this would help him intuit problems or inconsistencies faster than he could work them out precisely on paper.

I think this kind of concrete example is qualitatively different from an analogy or other way of visualizing the full system. The former helps more in specific circumstances, while the latter gives a warm, happy, and likely incorrect feeling of understanding.