Better Data Visualizations: A Guide for Scholars, Researchers, and Wonks: Summary & Key Insights

Before we choose a chart type, we must understand how people see and process visual information. Our perceptual system is attuned to differences—shape, size, color, position—and it seeks patterns and relationships naturally. In the book, I unpack how perceptual principles shape interpretation, referring often to foundational studies such as those by Cleveland and McGill, who demonstrated the accuracy hierarchy among visual encodings. Position along a common scale is most accurate, followed by length, angle, and area. That’s why a bar chart communicates a comparison more reliably than a pie chart.

When we remember that visualization is first a cognitive function, our design choices become purposeful rather than decorative. If your audience can’t readily perceive the relationship you wish to communicate, even the most sophisticated tool won’t compensate. Through examples drawn from policy datasets and research visuals, I demonstrate how our visual senses favor simplicity and directness, and how clutter, excessive shading, or unnecessary gridlines interfere with comprehension.

This section encourages you to think perceptually—to frame every design decision not merely as aesthetic but as communicative. Your charts should reveal the truth in the data by matching the right visual channel to the message you want to send.

The heart of the book is a tour through more than five dozen chart types—from humble bar charts to sophisticated network graphs. Each visualization type is examined by its purpose: comparison, distribution, relationship, composition, or flow. As I explain, no chart is inherently good or bad; appropriateness arises from the alignment between question and design.

Take the bar chart—a staple of business and academic communication. Its reliability stems from the human ability to compare aligned lengths easily. When differences are subtle, however, a dot plot may serve better, focusing the eye on position rather than column mass. When we track change over time, the line chart becomes indispensable. The narrative of a line chart lies in its movement: the slope tells a story of acceleration or decline. In contrast, scatterplots open our understanding to relationships—correlation, clusters, outliers—where context matters as much as the dots.

I devote considerable space to exploring the breadth of contemporary visualization: ridgeline plots for distributions, Sankey diagrams for flow, maps for spatial data, and network diagrams for relational data. Each type demands thoughtful alignment between purpose and viewer. Through these examples, I encourage readers to experiment but remain rooted in clarity—because novelty without precision is distraction, not communication.