Book Image

Graph Data Modeling in Python

By : Gary Hutson, Matt Jackson
Book Image

Graph Data Modeling in Python

By: Gary Hutson, Matt Jackson

Overview of this book

Graphs have become increasingly integral to powering the products and services we use in our daily lives, driving social media, online shopping recommendations, and even fraud detection. With this book, you’ll see how a good graph data model can help enhance efficiency and unlock hidden insights through complex network analysis. Graph Data Modeling in Python will guide you through designing, implementing, and harnessing a variety of graph data models using the popular open source Python libraries NetworkX and igraph. Following practical use cases and examples, you’ll find out how to design optimal graph models capable of supporting a wide range of queries and features. Moreover, you’ll seamlessly transition from traditional relational databases and tabular data to the dynamic world of graph data structures that allow powerful, path-based analyses. As well as learning how to manage a persistent graph database using Neo4j, you’ll also get to grips with adapting your network model to evolving data requirements. By the end of this book, you’ll be able to transform tabular data into powerful graph data models. In essence, you’ll build your knowledge from beginner to advanced-level practitioner in no time.
Table of Contents (16 chapters)
Part 1: Getting Started with Graph Data Modeling
Part 2: Making the Graph Transition
Part 3: Storing and Productionizing Graphs
Part 4: Graphing Like a Pro

Common igraph issues

While igraph is a powerful library for graph data and its analysis, much of its power comes from the fact that Python acts as an interface to igraph’s implementation behind the scenes, which is in C. Because of this, igraph comes with some quirks, which you need to bear in mind when using its features. In the following subsections we will take you through some common problems users come across when working with igraph in Python.

No nodes in the graph

Some graph and network science libraries in Python can be created directly from a list of tuples, an edgelist. In igraph, because of how node indexing is implemented in C, nodes have to be added to a Graph object before edges can connect them together. In the next steps, we will demonstrate this problem and the subsequent errors that arise:

  1. We can first read in edge data from musae_git_edges.csv as our example graph to load into igraph, using the inbuilt csv module. This file has headers, so we...