Use NetworkX to create graphs and use Matplotlib or Gephi to show it

Abstract

In this article I will simply introduce:

• NetworkX, use it to create a graph
• Matplotlib & Gephi, show the graph we created

The environment I use:

• Windows 10
• Python 3.6
• networkx 2.2 (Python Package)
• matplotlib 3.0.1 (Python Package)
• Gephi 0.9.2 (Under Java 11.0.1)

Create a Graph

First we download and install python package networkx, then import it:

import netwrokx as nx

next we create a graph with some nodes and edges:

G = nx.Graph()
G.add_node('a')
G.add_node('b')	# add node one by one
G.add_nodes_from([1, 2, 3, ])	# use list to add nodes
G.add_edge('a', 'b')		# add edge one by one
G.add_edges_from([('a', 1), ('a', 2), ('a', 3), ])	# use list to add edges

now we have a graph G with four nodes and four edges. We could create a directed graph use nx.DiGraph() the same.

Show it with Matplotlib

Now we can show this graph, first import python package matplotlib as followings:

import matplotlib.pyplot as plt

then we draw it and show it:

nx.draw(G, with_labels=True)
plt.show()

now check what we got:

LoL, marvelous, isn’t it?

Save the graph

We could save the graph we created into a file, we got many standad graph formats to choose, like GEXF, GraphML, Pickle etc. You can click here to see more formats of graph.
Now we have a python object or graph G, then save or read it using a pickle file:

nx.write_gpickle(G, 'the_great_graph.pkl')
F = nx.read_gpickle('the_great_graph.pkl')

NetworkX is a powerful tool to manipulate complex graphs, plenty of functions included, see the documentation to learn more what you need.

Show it with Gephi

What is Gephi?

Gephi is the leading visualization and exploration software for all kinds of graphs and networks. Gephi is open-source and free.
Runs on Windows, Mac OS X and Linux.

Some times we got a large graph maybe containing more than thousands nodes and edges then hard to show it with matplotlib, but we got another option to visualize it: Gephi.

Download & Install Gephi

Go to the official site to download it, the installation is very easy, just click next next and next.

Download & Install Java

Seems Gephi need Java 1.8 or later to open it, go to theJave official site download & install it, same easy, but remember the installation path, like C:\Program Files\Java\jdk-11.0.1.
next we go to the Gephi installation directory, then in the etc directory we could see the file gephi.conf, open it with any Editor like nodepad.
Find a line start with # jdkhome="..."', then move the character #, substitute the path in quote with Java installation path, like this:

jdkhome="C:\Program Files\Java\jdk-11.0.1"

Finally we save the gephi.conf file, then open Gephi.

Deal with large graphs

In Gephi, we can open a very large graph, and employ many Layout Algorithms to show the graph, like ForceAtlas 2.
Following is an example, this directed graph have 911,714 nodes and 1,216,998 edges:

References

[1] Overview of NetworkX. (2018, September 19). Retrieved from https://networkx.github.io/documentation/stable/index.html