Book Image

Network Science with Python and NetworkX Quick Start Guide

By : Edward L. Platt
Book Image

Network Science with Python and NetworkX Quick Start Guide

By: Edward L. Platt

Overview of this book

NetworkX is a leading free and open source package used for network science with the Python programming language. NetworkX can track properties of individuals and relationships, find communities, analyze resilience, detect key network locations, and perform a wide range of important tasks. With the recent release of version 2, NetworkX has been updated to be more powerful and easy to use. If you’re a data scientist, engineer, or computational social scientist, this book will guide you in using the Python programming language to gain insights into real-world networks. Starting with the fundamentals, you’ll be introduced to the core concepts of network science, along with examples that use real-world data and Python code. This book will introduce you to theoretical concepts such as scale-free and small-world networks, centrality measures, and agent-based modeling. You’ll also be able to look for scale-free networks in real data and visualize a network using circular, directed, and shell layouts. By the end of this book, you’ll be able to choose appropriate network representations, use NetworkX to build and characterize networks, and uncover insights while working with real-world systems.
Table of Contents (15 chapters)

The force-directed layout

The force-directed layout is a great visualization for many networks, and is a good go-to for your first visualization of a network. It works by repeatedly pushing all nodes apart and then pulling connected nodes back toward each other. It is benefits include the following:

  • Accommodates large networks
  • Clearly conveys community structure

However, the force-directed layout is also one of the most hairball-prone methods, particularly if there is one large community pulling all nodes together. It works best in sparser networks with multiple communities.

In Chapter 3, From Data to Networks, we saw that a default force-directed layout of the Frankenstein word co-occurrence network wasn't particularly informative. Now, we'll return to that example to demonstrate ways to focus on different aspects of a network and reduce clutter. The code to load...