Cool visualization, but the model needs work. The closest book to "Harry Potter and The Chamber of Secrets" is "The Victorian Lady's Guide to Sex, Marriage, and Manners".
I have a theory (after having searched for The Diamond Age) that certain very popular books are ironically not going to be close to similar books because they appear on so many varied reading lists.
There's probably a graph theory phenomena that describes what I'm thinking.
Within our data books like lord of the rings and harry potter are nearly impossible to map for "books like" because they are connected to so many other things. I am working now to fine tune our model, but it has been an interesting challenge.
Yeah basically what I think you're saying is that in a weighted graph, if there are edge weights which are orders of magnitude larger then the average it throws off certain models. Like nearest neighbor.
Basically just prune the top and bottom %1 of weighted edges to get an appropriate average. Would be my guess for a fix.
IIRC you could do pretty okay in Netflix's recommendation contest by ignoring what an individual likes and just recommending the stuff that "everybody" likes.
Hypothesis: the sort of person who reads "The Victorian...", is likely to both like Harry Potter books, and also to review them enthusiastically online. The typical HP child reader, does not review books online. Just an hypothesis.
There's probably a graph theory phenomena that describes what I'm thinking.