Introduction to NetworkX: Installation, Basic Usage, and Graph Operations in Python

NetworkX is a Python package for creating, manipulating, and studying the structure, dynamics, and functions of complex networks. It supports loading and storing graphs in various formats, generating random or classic graphs, analyzing structures, building models, designing algorithms, and visualizing networks.

Installation : While NetworkX can be installed alongside many related Python libraries, the easiest method is to install the Anaconda distribution, which includes NetworkX and many other useful packages for both Windows and Linux.

Basic usage (examples were run on Ubuntu with Visual Studio Code):

Creating an empty graph: <code>import networkx as nx G = nx.Graph()</code> In NetworkX, nodes can be any hashable object (strings, images, XML objects, other graphs, etc.), but None cannot be used as a node.

Adding nodes: <code>G.add_node(1)</code> From a list: <code>G.add_nodes_from([2, 3])</code> From an iterable nbunch (e.g., another graph): <code>H = nx.path_graph(10) # a path graph with 10 nodes G.add_nodes_from(H) # or G.add_nodes_from(H.nodes())</code> Adding a graph as a node: <code>G.add_node(H)</code>

Adding edges: <code>G.add_edge(1, 2) # equivalent to e = (1, 2) G.add_edge(*e)</code> Adding multiple edges from a list: <code>G.add_edges_from([(1,2), (1,3)])</code> Adding edges from an ebunch (any iterable of edge tuples, optionally with weight dictionaries): <code>G.add_edges_from(H.edges()) # cannot use G.add_edges_from(H)</code> Removing elements: <code>G.remove_node(H) G.clear() # removes all nodes and edges</code> Querying size: <code>print(G.number_of_nodes()) print(G.number_of_edges())</code>

Directed graphs and multigraphs: <code>DG = nx.DiGraph() # convert to undirected H = DG.to_undirected() # or H = nx.Graph(DG)</code> NetworkX provides MultiGraph and MultiDiGraph classes that allow multiple edges between the same pair of nodes, each possibly with different attributes.

Graph generators: <code>petersen = nx.petersen_graph() K_5 = nx.complete_graph(5) er = nx.erdos_renyi_graph(100, 0.15) </code> Reading and writing graph files (e.g., GML): <code>nx.write_gml(red, "path.to.file") mygraph = nx.read_gml("path.to.file")</code>

Analyzing graphs: <code>G = nx.Graph() G.add_edges_from([(1,2), (1,3)]) G.add_node("spam") print(list(nx.connected_components(G))) print(sorted(nx.degree(G).values())) print(nx.clustering(G)) </code>

Attributes: Nodes, edges, and the graph itself can store arbitrary key‑value pairs (e.g., weight , label , color ) via add_node , add_edge , or direct dictionary access such as G.nodes[node]["attr"] and G.edges[edge]["attr"] .

Visualization with Matplotlib: <code>import matplotlib.pyplot as plt nx.draw(G) plt.show() # or save to file nx.draw(G) plt.savefig("path.png") </code>

The basic introduction and usage of NetworkX end here; further topics such as advanced algorithms and integration with complex network research will be covered in future articles.