Ed Summers has done some nice analysis of the top hosts referenced in article space, based
on SQL dumps:
http://inkdroid.org/journal/2010/08/25/top-hosts-referenced-in-wikipedia-pa…
People with more in-depth knowledge might make something of this -- for instance the
importance of bots in external links, or the prevalence of certain types of information.
For instance, why/where/how is
edwardbetts.com used? (doesn't seem to be postcode
data, which was my first guess)
See also his linkypedia code:
http://github.com/edsu/linkypedia
-Jodi