Getting started with community detection

Before we can start, we need a network to use. Let’s use NetworkX’s Les Miserables graph that we used in the previous chapter since it held several separate communities:

1. Loading the network is simple:

   import networkx as nx

   import pandas as pd

   G = nx.les_miserables_graph()

That’s all it takes to load the graph.

2. There is a weight attribute that I do not want to include in the network because we don’t need edge weights for this simple demonstration. So, I’m going to drop it and rebuild the graph:

   df = nx.to_pandas_edgelist(G)[['source', 'target']]

   # dropping 'weight'

   G = nx.from_pandas_edgelist(df)

In those two steps, we converted the Les Miserables graph into a pandas edge list, and we kept only the source and target fields, effectively dropping the weight field. Let’s see how many nodes and edges exist in the network:

nx.info(G)
'Graph with...