I’d like to plug my paper, Exploratory data analysis for complex models (with a friendly discussion by Andreas Buja (who originally solicited the paper for the journal) and a rejoinder by me).

I think this paper will be of interest to many of you because (a) people respond positively to statistical graphics, but (b) generally think of graphs as methods for displaying raw data.

In my experience, graphs of statistical inferences–estimates from models–can be extremely useful.