Wayne Graham, in a post titled RSS Information Visualization, describes a Java applet he developed to show, in visual fashion, the links between blog tags and content. See the graph of his blog (it will take a few moments for the applet to load — be patient).
For his blog, each article is linked to each category in which it appears. From the map that’s generated, it’s easy to see that he writes mostly about XML and Cold Fusion, but also covers a wide range of topics. He describes the method he uses to generate the map in his posting.
A couple library-specific uses for this sort of tool come to mind. First is a book review blog — with genre or subject area as the tags, and individual reviews as the blog posts. It would be easy for a patron to ‘surf’ the map, looking for books in a subject. More often reviewed books could show up more prominently (other mapping technologies I’ve seen draw heavier lines depending on the strength or frequency of the relationship). Another would be even bigger — the library catalog itself, with subject headings (or, for bookstores like Amazon that have enabled tagging, user tags) as the subjects and items linked from there. This would be, as Wayne notes, a more sophisticated version of the tag cloud we’re all all becoming familiar with.