CODEWITHSUNDEEP
pythonnetworkx
simjas
simjas

Reputation: 43

to show the weights of edges in a projected network

I have projected a bipartite graph and made a new weighted graph. I would like to draw the graph, and show the edge weights.

Here is my attempt.

import networkx as nx
import matplotlib.pyplot as plt
from networkx.algorithms import bipartite
g=nx.Graph()
g.add_nodes_from(['s1','s2','s3','s4','s5'],bipartite=0)
g.add_nodes_from(['t1','t2','t3','t4'],bipartite=1)
g.add_edges_from([('s1','t1'),('s1','t4'),('s2','t1'),('s2','t2'),('s3','t1'),('s3','t4'),('s4','t3'),('s5','t2'),('s5','t3')])
l=bipartite.weighted_projected_graph(g,['s1','s2','s3','s4','s5'])
nx.draw(l, with_labels=True)
plt.show()

The node labels appear, but not the edge weights. How can I show the edge weights?

Upvotes: 4

Views: 3310

Answers (2)

yunmei huang
yunmei huang

Reputation: 11

I use this sentence that help me get weights: 

l=bipartite.weighted_projected_graph(g3,movies)

input: l.edges(data=True) output:[('The Dark Knight', 'The Matrix', {'weight': 1}), ('The Dark Knight', 'The Shawshank Redemption', {'weight': 1}), ('The Social Network', 'The Matrix', {'weight': 1})]

Upvotes: 1

Joel
Joel

Reputation: 23907

I've modified your code to include nx.draw_networkx_edge_labels

import networkx as nx
import matplotlib.pyplot as plt
from networkx.algorithms import bipartite
g=nx.Graph()
g.add_nodes_from(['s1','s2','s3','s4','s5'],bipartite=0)
g.add_nodes_from(['t1','t2','t3','t4'],bipartite=1)
g.add_edges_from([('s1','t1'),('s1','t4'),('s2','t1'),('s2','t2'),('s3','t1'),('s3','t4'),('s4','t3'),('s5','t2'),('s5','t3')])
l=bipartite.weighted_projected_graph(g,['s1','s2','s3','s4','s5'])
pos = nx.spring_layout(l)
nx.draw(l, pos = pos, with_labels=True)
nx.draw_networkx_edge_labels(l, pos)

This will probably show more than you want. I think it's set up for cases where edges may have somewhat arbitrary attributes. I think what you want is to do everything up to defining pos and then:

edge_weights = {(u,v,):d['weight'] for u,v,d in l.edges(data=True)}
nx.draw(l, pos = pos, with_labels=True)
nx.draw_networkx_edge_labels(l,pos,edge_labels=edge_weights)

Upvotes: 6

Related Questions