Complex Diagrams

1

I was recently asked some questions about the Beautiful Visualization (O’Reilly 2010) and my role as the technical editor and chapter contributor.

How did you end up working on Beautiful Visualization?
I was given the opportunity to work on the book because of my previous research and master’s thesis on methods of creating quality information visualizations.

Why is this book especially important now?
This is a particularly exciting time to be working with information visualization.

Visualization has become popular over the last few years. There have been some very good visualizations making it into the media and pop culture recently, and they have reached millions of people. Of note, the 2008 elections and current World Cup tournament have inspired dozens of visualizations that have received a lot of attention. Good visualizations are fun, educational, and engaging. People enjoy them, and some publications such as the New York Times and GOOD magazine are becoming known for their (generally high quality) work with information visualizations.

While visualizations are growing in pop-culture popularity, industry is also recognizing the power of visuals for analyzing and presenting huge sets of data. The amount of data that is being accumulated is always increasing, and good methods to extract actionable knowledge from that data are very valuable. The visual system is the highest-bandwidth channel into our brain, and it includes extremely sophisticated pattern recognition and pattern matching techniques. Trends, correlations, and outlying values become immediately apparent in a visual, where they might be overlooked in a spreadsheet.

In response to this growing popularity and demand for better visuals, many good tools and toolkits for creating visualizations have been released over the last few years. They allow people to create static and dynamic, interactive data visualizations. They are available for many languages, including JavaScript, Java, Flash/Actionscript, and Python. Protovis, Processing, Prefuse, and Nodebox are just a few of these toolkits.

Our book isn’t about specific tools, it’s about how people have designed good visualizations to satisfy the needs of their situation and audience. The need for these good examples and lessons continues to increase and more people are creating their own visuals.

What is the single most important thing readers of your book will be able to do after reading Beautiful Visualization?
Many chapters address different aspects of the design of visuals, and discuss why specific design choices were made. My hope is that our readers will become conscious of the choices they make when designing visualizations, and will be able to make constructive and useful design choices.

Who is your intended audience?
Primarily our audience consists of researchers, scientists, and other technical people with a need to analyze or visually represent data. However, our content will be interesting and useful to anyone with an interest in the mechanics of creating successful visualizations.

How important is the subject matter of your book? What do you think is on the horizon for your readers?
Currently, most visualizations are usually created by either graphic designers, who are good with aesthetics, or statisticians and programmers, who are good with code and numbers. Typically neither of these groups of people are trained in considering the needs of a specific audience. Ideally, in the future, visualization creators will be able to integrate technical, aesthetic, and usability perspectives on visualization. Our book is a step in this direction.

Can you suggest tips, tricks, and/or best practices for people interested in creating visualizations?
As with any creative endeavor, most of the rules of writing and design (think first, implement second) are broadly applicable. Here’s the entire process for creating a visualization, more or less.
1- Understand your own goals. Are you analyzing or presenting information? Once or in an ongoing way? What does this visual need to achieve? Et cetera.
2- Understand the needs, biases, and existing knowledge of your audience (users, customers, investors, etc.); their perspectives are likely different from yours. Your success is measured by their success.
3- Only include the necessary information; if it’s not helping get the message across it’s noise.
4- Define your axes so that placement becomes meaningful. This conveys a huge amount of information with very little ink.
5- Consider conventions and standards of your audience and how things occur in the physical world. Follow those conventions when reasonable to do so. Depart from them only when you can create a genuinely better format.
6- Be consistent with visual encoding, placement, and other recurring visual elements.
7- Iterate, test, iterate some more.
8- Don’t try too hard to be sexy; function first, sexy second.

[Crossposted at WeVisualize.info]