NetworkX and Matplotlib Graph Tutorial

NetworkX and Matplotlib Graph Tutorial | Generated by AI

Home 2025.10

Installing the Libraries

To use NetworkX and Matplotlib for drawing diagrams, install them if needed:

NetworkX: Handles graph creation and manipulation.
Matplotlib: Provides plotting capabilities.

Install via pip:

pip install networkx matplotlib

For faster rendering, Matplotlib recommends installing NumPy (pip install numpy).

Basic Setup and Example

Import libraries:

import networkx as nx
import matplotlib.pyplot as plt

Create a graph:
Use NetworkX to build a simple undirected graph (you can customize with directed graphs, weighted edges, etc.):

G = nx.Graph()  # Create an empty graph
G.add_nodes_from([1, 2, 3, 4, 5])  # Add nodes (e.g., 5 nodes)
G.add_edges_from([(1, 2), (2, 3), (3, 4), (4, 5), (5, 1)])  # Add edges (connections)

Draw and display the diagram:
Use Matplotlib’s backend for visualization:

nx.draw(G, with_labels=True)  # Draw the graph with node labels
plt.show()  # Display the plot

This produces a basic circular layout of the graph (a cycle with 5 nodes).

Advanced Customizations

Layouts: Control node positions (e.g., random, spring layout):

pos = nx.spring_layout(G)  # Force-directed layout for realistic networks
nx.draw(G, pos=pos, with_labels=True, node_color='lightblue', edge_color='gray')
plt.title("Network Diagram")
plt.show()

Node and edge styles: Customize appearance:

nx.draw(G, pos=pos, node_size=500, node_color='red', edge_width=2, font_size=10)

Directed graphs: For hierarchical relationships, use nx.DiGraph.

DG = nx.DiGraph()
DG.add_edges_from([('A', 'B'), ('B', 'C')])
nx.draw(DG, with_labels=True, arrows=True)
plt.show()

Saving the diagram: Export to file instead of displaying:
```
plt.savefig('diagram.png')
```

Tips and Resources

NetworkX docs (networkx.org) for graph types like trees or grids.
Matplotlib docs (matplotlib.org) for embedding in GUIs or subplots.
For large graphs (>100 nodes), use nx.draw_networkx or external tools like Graphviz for better performance.
Experiment in a Jupyter notebook for interactive plotting. If issues arise, common errors stem from missing backends (install via pip install pyqt5 or similar for interactive windows).

Back

x-ai/grok-code-fast-1

Donate