NetworkX package using dataframe with weights labeled

Question

I have a portion of dataframe as below (con_df):

from  to    fw
1     2    no fw
1     4    fw
2     3    no fw
2     5    no fw
2     6    no fw
2     7    fw

with 3 variables

want to make a graph using NetworkX

network = nx.from_pandas_edgelist(df=con_df, source='from', target='to',
                                  edge_attr='fw', create_using=nx.MultiGraph())
nx.draw(network, with_labels = True)

above is the code i used, the output is below, the problem is that the weight option is not shown on the graph. i thought that the edge_attr option is for the weight attribute, but it does not work

what i want is the fw variable is being labeled in the edge of the graph

is there a problem in the code or should i approach in different way?

Answer 1

with_labels refers to the nodes labels, not that of the edges.

You need to use nx.draw_networkx_edge_labels:

labels = {x[:2]: network.get_edge_data(*x)['fw']  for x in network.edges}
pos = nx.spring_layout(network)
nx.draw(network, pos, with_labels=True)
nx.draw_networkx_edge_labels(network, pos, edge_labels=labels)

output:

content of the labels dictionary:

{(1, 2): 'no fw',
 (1, 4): 'fw',
 (2, 3): 'no fw',
 (2, 5): 'no fw',
 (2, 6): 'no fw',
 (2, 7): 'fw'}