Session Ready
Exercise

Visualize similarity

The whole Madrid network can be difficult to reason about. One useful way to make it more comprehensible is to think about clusters of similar people. By filtering the similarity matrix, then converting it to a network, you can see how many group the whole network contains.

In the Madrid network, clusters of similar nodes correspond to terrorist cells. Can you spot them? We will investigate similarity between clusters deeper in the next chapter.

The similarity data frame sim_joined is loaded in the workspace for your convenience.

Instructions
100 XP
  • Filter the similarity data frame sim_joined for similarities greater than 0.6, assigning to sim_filtered.
  • Use graph_from_data_frame() to build a similarity network called filtered_network from sim_filtered.
  • Use ggraph() to plotfiltered_network with the Kamada-Kawai layout.
  • Add an edge link geom, mapping alpha to similarity.