Is there a way to group nodes in a Networkx graph together based on how connected they are?

Question

I have 39 people who have listed preferences for who they would like to be in a group with. I have input this data into a NetworkX graph in Python. I want to use the NetworkX graph to create groups of three based off how people are connected to eachother.

Answer 1

You are trying to identify communities. Here is a possible solution ( G is your graph, built using the networkx module):

from networkx.algorithms.community import greedy_modularity_communities

c = list(greedy_modularity_communities(G))

You can get very different outputs (ie, people can be divided in communities in different ways). This depends on the algorithm that you choose.