A Hierarchical Cluster Analysis of the Book Review Index

I recently produced a hierarchical cluster analysis of the Book Review Index data that I’ve been working with, though unfortunately I didn’t have enough time to include it in my ACH presentation last month. That’s a shame, since the resulting dendrogram is so interesting. You can view the visualization here. It’s too gigantic to comfortably embed on this page, but most browsers will let you open the image itself in a new tab.

The diagram itself was produced by first generating a similarity matrix comparing each of the 285 journals to each other journal. The underlying data is simply a collection of binary vectors indicating which journals reviewed each book. As such, I used the Jaccard distance as my similarity metric: the similarity of two journals equals the intersection of the books they both reviewed divided by the union of all books reviewed by either journal.