Community detection in NetworkX

NetworkX provides several community detection algorithms. There are many ways to define communities, and many ways to search for them. However, in general, the NetworkX community detection functions take a Graph object and return a list or iterator over communities, represented as a set of node IDs.

The simplest type of community detection searches for non-overlapping communities. In other words, a network is partitioned into communities such that each node belongs to exactly one community.

Modularity maximization

One approach to finding such a partition is to define a function that quantifies the quality of a set of partitions, and then adjust the partition to maximize the quality. One popular...