Hi! I am using igraph (Python) to calculate betweenness centrality and was surprised by how fast it was – the betweeness centrality calculation for ~40k nodes and ~50k edges takes about 3 minutes, while it never finishes when I try to use networkx. I’m curious what algorithm igraph uses, and how is it so much faster?
For completeness, betweenness calculation is implemented here:
@alyssa: If you like, you are welcome to submit a pull request that includes a reference to Brandes’s paper in the documentation. It’s probably good to mention it there.