I have a link base data which I did clustering label for each of them and the data look like this: the id_1 and id_2 are linked to each others and the label is the cluster number for each link.
id_1,id_2,label
0,18427,1
1,21708,1
1,22208,1
1,22171,1
1,6829,1
1,16590,1
1,20135,1
1,8894,1
1,15785,1
1,10281,1
etc..
and I use networkx library to graph the line between them.
Is there a way to color the data Nodes based on cluster label as we do in the seaborn library?
nx.draw has plenty of arguments that allow you to customize the generated plot. In this case you want to set a edge_color that depends on the label attribute. I've modified the labels a bit here to better illustrate the point.
To reproduce your example I've generated the graph from the edgelist as a pandas dataframe, which makes it very simple. Since I've set edge_attr=True, the label will be an attribute. We can thus save it as a dictionary using nx.get_edge_attributes.
Then we can set the edge color using edge_color=list(labels.values()) with a cmap of choice and plot the graph as follows:
import matplotlib.cm as cm
from matplotlib import pyplot as plt
G = nx.from_pandas_edgelist(df, source='id_1', target='id_2', edge_attr=True)
cmap = cm.get_cmap('viridis', max(labels.values()))
labels = nx.get_edge_attributes(G, 'label')
plt.figure(figsize=(8,6))
nx.draw(G, with_labels=True,
edgelist=list(labels.keys()),
edge_color=list(labels.values()),
edge_cmap= plt.cm.summer,
node_color='lightgreen',
node_size=1000, width=2)
Set up -
s = StringIO('''
id_1,id_2,label
0,18427,1
1,21708,1
1,22208,3
1,22171,4
1,6829,1
1,16590,2
1,20135,3
1,8894,2
1,15785,2
1,10281,2
''')
df = pd.read_csv(s, delim_whitespace=False, sep=',')