A correlation network is a network chart made from a correlation matrix. Here we take the example of the mtcars dataset. For 32 cars we have several features. We can compute the correlation between each pair of car using the cor function:

This is an adjacency matrix (square) that we can give to Igraph using the graph_from_adjacency_matrix() function. (see graph xxx). Note that we keep only connections stronger than 0.995.

Now it is easy to compute the default network chart

But it is more interesting to map a color on it and add a few customizations:

Not what you are looking for ? Make a new search !


Leave a Reply

Notify of