Interview

Off the Charts: An Interview with Scott Murray, Code Artist

Posted by dave on April 3, 2014 Off The Charts

Design isn’t beautification; it’s constructing systems.

Recently I had lunch with Scott Murray and took the opportunity to interview him for our Off the Charts series. Scott is an Assistant Professor of Design at USF, the author of Interactive Data Visualization for the Web , a contributor to Processing and a “code artist” with an enormous body of work. He’s done so many interesting things that it is difficult to cover him in a single interview but I’ve done my best.

You are a professor, writer, speaker, developer, artist and many other things. On your website you sum all of that up with the title “code artist”. What does that mean?

Some of us working with art, design, and technology really struggle with titles. I’ve met creative technologists, data artists, creative coders, visualizers, mapmakers, and installation artists. But with an interdisciplinary practice, it can be tricky finding the right language for describing your work to people outside the field. I chose “code artist” because I felt code was the only medium all my work has in common, whether that work ends up in print, on the web, or as a projected image. “Artist” ultimately made more sense than “designer” because “design” implies a specificity of communication. My background is in design, and I teach design courses, but “artist” feels like a more flexible term.

Of course, as the language around this practice evolves, I may change my own title. (As an artist, you do things like that.)

I chose “code artist” because I felt code was the only medium all my work has in common, whether that work ends up in print, on the web, or as a projected image.

You wrote a book, Interactive Data Visualization for the Web. Who is the book targeted to?

The book is an introduction to D3 for people with little or no web programming experience. It’s really for designers who want to get into programming, or programmers who want to get into data, or journalists and analysts who are used to working with information, just not in the context of visuals on the web.

1.jpg

As a design professor, why did you decide to write a book about a programming framework like D3?

With D3’s initial release a few years ago, it was clear that it was going to be a powerful, important tool for working with data on the web. But its conceptual model is quite different than most other tools. I found myself wanting to work with it, but having difficulties with the learning process. So I took notes on everything I found challenging about it, and then wrote down how I solved problems as they arose. That list became the basis of the tutorials on my site, which I later expanded greatly into a full book. I figured since I already struggled through the process, perhaps I could make the process a bit easier for others.

When do you recommend using D3? And how do you recommend people learn more about D3?

D3 is ideal for two-dimensional data graphics on the web. It’s really the best tool for charts, graphs, and maps on the web. If you’re doing 3D, or something for another context outside of the browser (like print media), then there may be other tools better suited to your task (like Processing). Obviously I hope people will check out my book. It’s a great starting place, and the first chapters include lots of pointers to other resources and non-D3 tools. It should help you decide whether or not D3 is right for your project.

What type of classes do you teach at USF?

I teach some of our foundational courses - introductions to visual communication for static media (e.g., paper) as well as dynamic media (e.g., web) - but also some upper division electives, like Information Visualization and Interaction Design.

What is the number one insight you want your students to take away from these classes?

Design isn’t beautification; it’s constructing systems.

You worked on Kindred Britain. Tell us about that project and the visualization considerations that went into it.

There were too many considerations to list here, but three of them were color, color, and color! KB has three main visualization panels: a network view, a timeline view, and a geographic map. All three views are linked, so, for example, any people appearing in the network panel also appear in the timeline, and the geographic locations of their life events are indicated in the map. With so much information displayed in a single window, we fought information overload by eliminating all color from the interface. With menus and navigation entirely grayscale (except for links), any color really pops out, even in small doses. We wanted to make sure that the presence of color indicated the presence of information. Everything else should be black, white, or gray.

Read more of Scott’s considerations here: Kindred Britain.

Can you discuss two of your favorite visualizations? Either ones you have done or ones that you admire.

I love Close Votes by Jan Willem Tulp. It’s an interactive graphic (made with D3) that visualizes similarities in voting patterns across cities in The Netherlands. It’s visually beautiful, and the different layouts make it quite easy to identify patterns and outliers in the data - even though I know nothing about Dutch politics!

data art1

data art2

On the art side of things, unnamed soundsculpture by Daniel Franke and Cedric Kiefer is a stunning work. They used Processing to collect data from three Kinects, each of which was “filming” a dancer from a different position. The points were then merged into a single cloud, and rendered as video. It’s an abstract, meditative flow of digital particles that originated from a human form. I recommend watching the “making of” video.

When you land on your website, there are beautiful particles. Was code art involved in that? How did you go about making that?

It’s up to the visitor to decide whether or not those circles qualify as art. I was just trying to design something engaging to fill the otherwise empty space. Particles are fun because our brains can only track around four or five moving objects at once. Beyond that, the effect is a bit overstimulating, but in a calming way, as we get lost in the larger texture of motion, like watching waves in the ocean.

On the technical side, I had an idea of what I wanted to make, and started sketching using Processing. The final piece on my website uses Processing.js, a JavaScript adaptation of Processing.

Particles are fun because our brains can only track around four or five moving objects at once. Beyond that, the effect is a bit overstimulating, but in a calming way, as we get lost in the larger texture of motion, like watching waves in the ocean.