Author Networks

The network graphs below show an individual author's similarity network. Authors are deemed "similar" based on Goodread's "Similar Authors" page for an author. For example, you can see that author Dante Alighieri is similar to other writers like Homer, Ovid and Montaigne. These similarity measures are based on Goodreads user behavior: users who read Dante also read Homer, Ovid and Montaigne. This may not be the best way to measure similarity, but it provides for access to a very wide and broad author network that we otherwise wouldn't have access to.

As these are the networks for an individual author, the only authors that are included are the following (using Dante as an example):

• authors that have Dante on their "Similar Authors" page
• authors that are included on Dante's "Similar Authors" page

Further, the edges (connections) are drawn whenever applicable. Naturally, all the authors in Dante's network will have a connection to him; further, some authors, like Homer and Ovid, have each other on their "Similar Authors" page, so a connection is drawn between them as well. You'll often notice a large cluster of nodes/authors all beside each other. These are authors that only have a connection to Dante, but no one else in Dante's network.

The Louvain method is used for community detection and node coloring. Node and label sizing are based on a node's degree (number of "similar" authors). Cutoffs for inclusion based on Goodreads counts were made in order to reduce the node/edge counts. The node sorting will automatically begin and stop after about thirty seconds. You can click the "Play" button to continue the animation.

source: Goodreads "Similar Authors"