generate adjacency matrix via networkx that has weights

Question

I'm using the code below to generate some random Euclidean graphs. I expected the adjacency matrix to have the distances between nodes rather than just being a binary representation of the connections. How do I make the adjacency matrix hold the actual distances/weights between nodes?

g = nx.generators.random_geometric_graph(n, 10)
adj = nx.adjacency_matrix(g).todense()

Paul Brodersen · Accepted Answer · 2022-04-04 10:23:16Z

1

Well, the adjacency matrix denotes the adjacency, not a distance. To get the distances, you can grab the pos attribute of the nodes, and proceed normally:

g = nx.generators.random_geometric_graph(10, 0.1)
pos = np.array([g.nodes[node]['pos'] for node in g.nodes])
delta = pos[np.newaxis, :, :] - pos[:, np.newaxis, :]
distances = = np.sqrt(np.sum(delta**2, axis=-1))

answered Apr 4, 2022 at 10:23

Paul Brodersen

12.7k26 silver badges51 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Brannon Over a year ago

If you were going to set the weight attributes on the edges, would you do it this same way?

Paul Brodersen Over a year ago

In that case, you may be only interested in the distances between connected nodes. If the graph is small, computing the distances between all nodes is fine, but if the graph is very large, you might want to only compute the distances between connected pairs of nodes (i.e. loop over edges, get the node positions, compute the distance, set edge weight, etc).

Collectives™ on Stack Overflow

generate adjacency matrix via networkx that has weights

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related