Below is a plot of the podcast network graph. Podcasts are linked if they share at least one guest in common, and the strength of the links are determined by the number of common guests and their total time spent on each podcast. This is a version of a weighted Bipartite Graph Projection. The podcast nodes are colored by the number of other podcasts they share guests with. The placement of the nodes on the graph is set using the Fruchterman-Reingold force-directed algorithm, which plots nodes so that all the edges are of more or less equal length and there are as few crossing edges as possible. This has the effect of pulling more central nodes towards the center and less central nodes to the periphery.
We can also plot a network graph of the people (hosts and guests) in the podcast network. Due to computational constraints, we cannot plot everyone in the network in one plot, so this plot only includes podcast hosts, and guests who have been on at least 3 different podcasts. This decreases the number of nodes from over 12,000 to under 1,300. We colored the nodes according to their assigned categories (see the categories page), and we related the size of the node to its Pagerank. We can see from this plot how the Fruchterman-Reingold force-directed algorithm pulled people of the same category into groups. This roughly matches what we see in the category mixing figure on the categories page.