Discover Communities: A Python Library for Graph Clustering Visualization

Recently a new Python visualization tool has become popular.

This tool, called communities , is a Python library for detecting and visualizing community structures in graph clustering problems.

Key Features

It supports multiple algorithms, including Louvain, Girvan‑Newman, hierarchical clustering, spectral clustering, and Bron‑Kerbosch, and can visualize the results.

Importing an algorithm and providing an adjacency matrix

Example using the Louvain algorithm. The algorithm works on an undirected graph represented by a 2‑D NumPy adjacency matrix.

import numpy as np
from communities.algorithms import louvain_method
adj_matrix = np.array([[0,1,1,0,0,0],
                      [1,0,1,0,0,0],
                      [1,1,0,1,0,0],
                      [0,0,1,0,1,1],
                      [0,0,0,1,0,1],
                      [0,0,0,1,1,0]])
communities, _ = louvain_method(adj_matrix)
# >>> [{0, 1, 2}, {3, 4, 5}]

After running the algorithm, the list of communities (each a set of node indices) is obtained.

Visualization with color coding

The draw_communities function visualizes the graph, assigns colors to each community, and allows options such as dark or light background, image resolution, and saving to file.

draw_communities(adj_matrix : numpy.ndarray,
                 communities : list,
                 dark : bool = False,
                 filename : str = None,
                 seed : int = 1)

Parameter meanings:

adj_matrix (numpy.ndarray): adjacency matrix of the graph.

dark (bool, default=False): dark background if True.

filename (str or None, default=None): path to save PNG; None shows interactively.

dpi (int or None, default=None): image resolution.

seed (int, default=2): random seed.

Animation of the algorithm

The louvain_animation function creates an animated GIF showing how nodes are assigned to communities over iterations.

louvain_animation(adj_matrix : numpy.ndarray,
                  frames : list,
                  dark : bool = False,
                  duration : int = 15,
                  filename : str = None,
                  dpi : int = None,
                  seed : int = 2)

Key arguments are similar to the visualization function, with frames containing per‑iteration community mappings and modularity values.

The library can also compute community adjacency matrices, Laplacian matrices, and modularity matrices.