Introduction

This book is about the application of graph visualization and analysis for business. Graph applications are a unique and valuable resource for discovering actionable insights in data. In recent years, analysts inside some of the world's most innovative companies have been intensively exploring graph-based approaches to a gain deeper understanding of the dynamics of their businesses while discovering opportunities and strategies for improvement.

As the volume, variety, and velocity of available data has grown, so has the need for techniques and technology to make sense of it all. Organizations have become acutely aware of the limitations of simple dashboard-style charts. Dashboards are good at showing metrics and trends. They can inform you when areas of business are underperforming or outperforming others, but they cannot begin to tell you why, and understanding why is key to taking effective action.

The function of a graph is to represent links between things, revealing the structure and nature of relationships in data. Relationships are fundamental to the why and the how of things, which is one of the reasons graph analysis and visualization has so much potential for value.

Looking back on 20 years of our personal history designing and building new applications for business and intelligence analysts, the authors realize that graphs have played a role in many of those solutions. Today, several of our most significant research and software development efforts are, in essence, graph-based.

Despite the utility of graphs, however, little has been published about the application of graphs outside of the world of science, and even less has been published about graph design. With recent advancements in the capabilities of open source graph tools and libraries, graphs have become accessible to every business analyst, but access to knowledge of effective principles and techniques for graph analysis and visualization remains relatively limited. Our hope in writing this book is to help change that.

Who This Book Is For

This book is for data scientists and analysts interested in applying graph analysis to decision-oriented problems. The examples provided are taken from the business world, but the principles and techniques used are highly relevant to government and non-profit problems as well.

No prior knowledge of graph theory or practice is required. A reader who is new to graph analysis should find it useful to read this book from start to finish. More experienced readers may choose to skip ahead to subjects of interest in Part 3, which expands in detail on specific analytic themes.

Some examples in this book include light programming, but the majority of sample applications use point-and-click tools. In both cases, a moderate level of technical aptitude will be required.

How This Book Is Structured

This book is composed of four parts. The first part represents a broad introduction to the subject of graphs. Subsequent parts are organized into progressively more specialized or advanced topics. Chapters 3 through 10 are written by Richard Brath, and the remaining chapters by David Jonker.

• Part 1-In the first part of the book, the authors provide an overview of graph applications in business and introduce various types of graphs, which are covered in more detail in Part 3.
• Part 2-The second part provides a comprehensive look at the major steps in the process of graph visualization and analysis.
• Part 3-The third part of this book is organized into distinct analytic themes and associated graph types and techniques.
• Part 4-The fourth part focuses on advanced topics representing areas of ongoing research, as well as fundamental design principles.

Materials for Download

This book includes online data files, source code distributions, and graph visualization files to accompany the examples provided. These Supplemental Materials are organized by chapter. The software required to view or run these files is described in each of the chapter examples. Files for download include the following:

• Data files-Most data files are available in a generic format such as text (.txt) or comma-separated values (.csv), which can be read directly into graph software or otherwise used by programs. In some cases, there will be two files, one for nodes and one for edges (that is, the links between nodes). In other cases, graph data files will be provided in a graph-specific file format, such as .gdf or .graphml. These are formats that many graph tools import directly.
• Excel files-There are a few Excel spreadsheet examples identified by .xls or .xlsx file extensions. These require Microsoft Excel in order to run.
• Graph visualization files-Some examples also include graph visualization files such as .gephi or .cys. These are files associated with specific graph visualization software such as Gephi or Cytoscape, respectively. To view these files, you must first download the free graph visualization software package and install it. See the following section for details.
• Python code-Programming examples use the Python language. These programming files are identified by the extension .py. Python examples are done in version Python 3.x and require the download and installation of Python. See the following section for details.
• HTML and JavaScript-Examples using JavaScript are typically web pages containing JavaScript and identified as .html files. These files will run in a standard modern web browser such as the latest version of Chrome or Firefox.

Source code for the samples is available for download from the following website:

www.wiley.com/go/GraphAnalysisVisualization

What You Need to TRY THE EXAMPLES

A variety of tools are used in the book to process data and/or visualize data. In order to use the data files previously identified, the following software may be required:

• Gephi-The end-user point-and-click free software product Gephi (https://gephi.github.io/) is used for many of the graph visualization examples in the book. Many of the data files can be imported into Gephi for analysis and visualization. Chapter 7 of the book discusses some of Gephi's features, building on the basic graph analysis process described in Chapters 3 through 6.
• Cytoscape-Cytoscape (www.cytoscape.org/index.html) is another free end-user software tool for graph analysis used in many examples in the book. Many of the data files can also be imported in Cytoscape for analysis and visualization. Chapter 7 discusses some of Cytoscape's features and also outlines some of the differences between Gephi and Cytoscape.
• yEd-yEd (www.yworks.com/en/products/yfiles/yed/) is an alternative free end-user point-and-click software product made by yWorks for graph analysis and visualization.
• Excel-Microsoft Excel (http://products.office.com/en-us/excel) spreadsheets are used in several examples. Excel is not free, but most readers will already have a copy, and Microsoft does allow download for time-limited evaluations. Several examples also use the NodeXL plug-in for Excel.
• NodeXL-Excel allows developers to create plug-ins that access and enhance Excel's functionality. NodelXL (http://nodexl.codeplex.com/) provides graph functionality for social network data retrieval, as well as graph analysis and visualization.
• Python-For programmatic manipulation of data, the Python 3 (https://www.python.org/) programming language is used in some examples. Python is freely available.
• A modern browser-While any modern web browser should be capable of viewing the JavaScript/HTML examples, Chrome (https://www.google.com/intl/en_us/chrome/browser/) was the browser used by the authors.
• D3.js-D3 (http://d3js.org/) is a JavaScript library used to create a variety of interactive data visualizations in a browser, and used, for example, in Chapter 8.
• Aperture JS-Aperture JS (http://aperturejs.com/) is a JavaScript framework library used in some of the examples in the later part of the book, for example, in Chapter 12.
• Titan-A Titan (http://thinkaurelius.github.io/titan/) graph database is used for several big data examples found in Chapter 14.

To use these software libraries and tools, you will need to download them yourself and install them, with the exception of the JavaScript libraries, D3.js, and Aperture JS. These are packaged with the examples for download from the companion website specified earlier.

Caveats

The chapters in this book use case study examples to illustrate various applications and forms of graphs and how to use them yourself. Illustrations make use of real tools and real data where possible. There are caveats to keep in mind with both of these.

While the authors have used open source tools that are freely available to anyone, many of these tools are still works in progress and, as such, lack some of the polish and robustness you might expect of a finished product. Expect that a little extra patience will, at times, be the price of being an early adopter. Another aspect of documenting work-in-progress tools is that they are more likely to change. Use the tool-related steps in this book as general guidelines to a process. If the...