Networkx: Graph Labels getting mixed up and not adjusted

Question

So I made my graph using networkx library which is a unipartite graph. When I put the labels it doesn't seem right, its getting mixed up. There are some words which has more length than the others and its out of the boundary of the nodes. Can my graph be adjusted so everything look clear and understandable?

And also I want the node to be like a dot and the labels appearing over the nodes not inside the nodes.

Like this

Graph I made..

Here is the code..

import networkx as nx
import matplotlib.pyplot as plt
import numpy as np
from networkx.algorithms import community
from networkx.algorithms import bipartite


G1 = nx.read_edgelist("abcd.txt")
print(nx.info(G1))
c = bipartite.color(G1)
nx.set_node_attributes(G1, c, 'bipartite')
type1  = {n for n, d in G1.nodes(data=True) if d['bipartite']==0}
type2  = {n for n, d in G1.nodes(data=True) if d['bipartite']==1}
G = bipartite.projected_graph(G1, type1)
type2g = bipartite.projected_graph(G1, type2)


pos = nx.spring_layout(G,k=0.30,iterations=50)

nx.is_bipartite(G1)
#average clustering
nx.average_clustering(G1)

#Diameter
print("Diameter:",nx.diameter(G1))

options = {
    'node_color': 'purple',
    'node_size': 40,
    'line_color': 'yellow',
    'linewidths': 0,
    'width': 0.3,
}

#degeree plotting
def plot_degree_distribution(wiki):
    degs = {}
    for n in wiki.nodes():
        deg = wiki.degree(n)

        if deg not in degs:
            degs[deg] = 0

        degs[deg] += 1
    items = sorted(degs.items())

    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot([k for (k, v) in items], [v for (k, v) in items])
    ax.set_xscale('log')
    ax.set_yscale('log')
    plt.title("Degree Distribution")
    fig.savefig("degree_distribution.png")


# plot_degree_distribution(G)
d = []  # create a set
for n in G.nodes():
    d.append(G.degree(n))

ec = []  # create a set
for e in G.edges():
    if (G.degree(e[0]) > G.degree(e[1])):
        ec.append(G.degree(e[0]))
    else:
        ec.append(G.degree(e[1]))

pos = nx.spring_layout(G,k=1.5, iterations=200)

factor = 25  # to change the size of nodes with respect to their degree


nx.draw_networkx(G, pos,
        edge_color=ec, edge_cmap=plt.cm.plasma,  # edge color
        node_color=d, cmap=plt.cm.plasma,  # node color
        node_size=[x * factor for x in d])  # node sizse
plt.savefig ("simple_graph.png")
fig = plt.gcf()
fig.set_size_inches((10,10))
plt.show()

TavoGLC · Accepted Answer · 2019-06-27T17:37:08.700

1

Changing the spring layout to a circular layout will help you to get a better label visualization.

cgraph=nx.gnp_random_graph(15,p=0.5)
labels={}
for k in range(15):
    labels[k]='long named node '+str(k)

circPos=nx.circular_layout(cgraph)
nx.draw_networkx_labels(cgraph,pos=circPos,labels=labels) 
nx.draw(cgraph,pos=circPos)

Also if the edge label is difficult to visualize, you can change the position of the labels by changing the layout position as follows.

pos_attrs = {}
for node, coords in circPos.items():
    pos_attrs[node] = (coords[0]+0.1*(-1)*np.sign(coords[0]), coords[1]+0.1*(-1)*np.sign(coords[1]))
nx.draw_networkx_labels(cgraph,pos=circPos,labels=labels) 
nx.draw(cgraph,pos=pos_attrs)

Another great way to deal with labels on circular layouts is presented here: NetworkX node labels relative position

Hope it helps

edited Jun 27 '19 at 17:37

answered Jun 27 '19 at 01:50

TavoGLC

889
11
14

pos_attr giving me the error 'float object not callable' – Abdul Wadood Jun 27 '19 at 07:14
i updated the answer, an * was missing over `coords[1]+0.1*(-1)*np.sign(coords[1])` – TavoGLC Jun 27 '19 at 17:38
It isn't working the labels are still appearing on the nodes – Abdul Wadood Jun 28 '19 at 15:47
i am using spring layout is it because of that? – Abdul Wadood Jun 28 '19 at 15:54
Yes, its because the layout. If you need to use that kind of layout, try to cluster the nodes into categories, or just add a label to the most important node. Hope it helps – TavoGLC Jun 28 '19 at 21:36

Networkx: Graph Labels getting mixed up and not adjusted

1 Answers1