Python for graph and network analysis is a short monograph (200 pages) in the “Advanced Information and Knowledge Processing” series. It is composed of eight chapters and two short appendices. It is described as suitable for courses on social network analysis (SNA), but the audience is not clearly defined.

The book starts with a chapter introducing the theoretical concepts of network analysis including the sociological meaning of network relations, measurements, distribution, segmentation, and recent developments in network analysis, and it ends with a short example using the R program iGraph.

Chapter 2 presents network basics and introduces NetworkX, software written in Python that will be the core tool used in the book. Snippets of Python code illustrate the concepts including a few aspects of matrices. I wish the book had a companion website where the complete code could be found in order to avoid typing the code to practice the examples.

Chapter 3 illustrates graph theory, its origins, and basics such as vertices and edges, and provides code for the main graph traversal options.

Chapter 4 introduces social networks and their properties. The rising importance of online social networks is notable, and the authors briefly mention some of the tools used in SNA modeling.

Chapter 5 concentrates on Ego network analysis, identifying influential individuals. Centrality is at the core of the chapter, and different attributes are evaluated such as degree, closeness, betweenness, and eigenvector. The authors cover other concepts such as PageRank, neighbors, and bridges.

Chapter 6 moves on to group-level analysis with an explanation of cohesive sub-groups, cliques, clusters, triadic analysis, structural holes, brokerage, transitivity, and other factors. It is the longest chapter of the book.

A short chapter 7 completes with network-level analysis, clarifying concepts such as components/isolates, core/periphery, density, shortest path, reciprocity, affiliation network, two-mode networks, and homophily.

Chapter 8 concludes the book with information diffusion in the social network. Diffusion and adoption are illustrated through a number of conceptual examples and limited sample code.

The remainder of the book includes two short appendices, one on Python syntax and the other providing a NetworkX tutorial. Neither would be sufficient to achieve any level of proficiency to help one code or extend the Python examples in the book.

While the topic is certainly interesting and worth studying, the lack of a defined audience is detrimental to achieving clear goals. As a research monograph, the book is missing extensive references and an index. It could be supportive of the work of a postgraduate student, but in itself this book does not offer a sufficient foundation for publishable research as it is not advanced enough. As an undergraduate textbook, the monograph lacks the introductory aspects, clarity, complete code, supporting website, and exercises necessary to support a course. The book is complementary to a reading list for someone who already has a good understanding of graph theory and social network analysis and has the wish to validate their view or is coming from another path such as social network analysis using R.

As a standalone basis for network analysis using Python, I am afraid that the book is not compelling.

More reviews about this item: Amazon