Making Data Visual – A Practical Guide to Using Visualization for Insight, by Danyel Fisher and Miriah Meyer is a slim volume that fills an important niche in the data visualization literature. It is not your typical data visualization book that provides tips on creating charts, the use of color, or the importance of annotation. Readers looking for a text on visual design may be disappointed, but readers interested in the larger visualization design process will find the book rewarding.
The early chapters and examples at the end are the most valuable part of the book. After a quick introduction in “Getting to an Effective Visualization,” the book moves to its principle contributions. “From Questions to Tasks” focuses on the importance of asking and clarifying questions in an iterative process. The authors note that “The process of breaking down questions into something that can actually be computed from data is iterative, exploratory, and sometimes surprising.”
Fisher and Meyer use the example of “Identifying Good Movie Directors” to make their point. They effectively use the example to show the importance of “making questions concrete.” To do this, they break the question into objects, measures, groupings (or partitions), and actions in an iterative process. They show how proxies, or related data that substitutes for the data of interest, may be used to answer questions. Their process includes defining a movie, a “good” movie, a “good” movie director, finding proxies, and understanding the relationship between movies and directors.
The approach is valuable, leading to new questions that fall into four categories: 1) revealing that a new analysis is needed, 2) identifying new research directions, 3) starting additional exploratory data analysis, or 4) surfacing data quality issues that require further attention.
In “Data Counseling, Exploration, and Prototyping,” the authors move on to techniques that support the process of going from questions to tasks. These include data counseling, exploration, and prototyping. Data counseling is an interactive communication process with stakeholders: to understand their intent regarding the data, to become familiar where the data comes from, and to identify any problems associated with the data. Exploring the data involves … well … exploratory data analysis and visualization. Rapid prototyping covers a range of activities from low-fidelity sketches through complex interactive visualizations. These differ in the amount of energy, commitment, and speed of iteration.
“Components of visualization” is a short chapter about data … the types of data, the transformations between data, and dimensionality reduction and clustering.
The next two chapters cover visualization and are divided into “Single Views” and “Multiple and Coordinated Views.” In “Single Views,” Fisher and Myer observe that this is “well-trodden territory.” Other books cover the topic in greater depth, but their brief discussion includes the usual suspects: bar charts, histograms, line charts, box plots, scatterplots, word clouds, and maps, as well as the lesser used link-node charts, adjacency matrices, treemaps, and sunbursts.
The “Multiple and Coordinated Views” chapter discusses small multiples, scatterplot matrices, multliform views and dashboards, overview + detail, and overlays. This content deals with multiple-linked views and interaction, so it may be less familiar to readers.
There are two case studies: “Case Study 1: Visualizing Telemetry to Improve Software” and “Case Study 2: Visualizing Biological Data.” The value of these chapters is their demonstration of the end-to-end process described throughout the book.
Case Study 1 deals with the creation of a visualization for Microsoft to explore improvements of new software releases versus the release they replace. Case Study 2 is a “somewhat more complex example” that shows how similar sets of genes produce different results across species. Case Study 2 may challenge the average reader.
All-in-all this book successfully fills an interesting gap in the data visualization literature, dealing with end-to-end visualization process. It covers the importance of asking good questions, exploring data, matching data to visualization techniques, and constantly iterating and refining the process. It is both thoughtful and clear in its approach. That said, it will likely disappoint readers who pick it up thinking they are getting a data visualization book about graphs, charts, and maps.