Visualization: Just Look What’s Coming

Today’s performance dashboards are cheap toys compared to the tools coming down the information visualization pipeline

Every now and then all of us have moments when we see something new and say, "This is big." Think back to when you received your first e-mail message or got your first glimpse of the Web. I had one of these moments two weeks ago at an information visualization conference.

Visualization’s ability to reveal stories hidden in plain sight was, pardon the pun, eye opening. The experimental new Web sites and research shown at the InfoVIS conference in Sacramento, Calif. made most of today’s performance dashboards look like cheap toys.

This wasn’t a business crowd. Several hundred grad students and geeks mostly in their 20s, 30s, and 40s from all over the world gathered in jeans, running shoes, and wrinkled cotton shirts eager to talk about algorithms. This year a healthy contingent from business came too, comprising about one third of attendees, up from nearly zero last year.

New York Times graphic editor Matt Ericson’s opening keynote showed information visualization—or infovis for short—at its most refined: the crisp infographics produced by his team of 30 editors, reporters, and cartographers.

One was an online map that showed where presidential candidates been seen lately (see and sources of funds. Users can filter by party, period, and candidate.

In another chart, he demonstrated context with Iraqi war deaths compared to past wars. Shown alone, the bar graph looks dramatic, against Vietnam deaths it seems less so, and World War II seems to tower over both.

It’s news experienced more directly than you can read in a text narrative.

Ericson told how an infographic changed Bill Gates’s mind. Several years ago, Gates told Times writer Nicholas Kristof how Gates and his wife had decided to fund the fight against AIDS instead of wiring Africa for the Internet. Kristof’s articles helped, of course, but what really grabbed him, he said, was a simple graphic listing third-world health problems and how many people they kill that ran alongside one of Kristof’s articles.

One of the most public of experiments is the YouTube of information visualization, Many Eyes, produced by IBM’s Visual Communication Lab. It lets anyone upload data and use standard templates to create enlightening displays.

Hearing former Attorney General Alberto Gonzales’ Senate testimony is one thing, but seeing an analysis of the text is also interesting. Fernanda Viegas, who co-founded the project with Martin Wattenberg, talked about one user’s a "word tree" analysis (see It reveals that the most common first words in sentences were "I don’t," and the most common third word was "recall." One long pattern even suggested a memorized response.

Within an hour of that testimony appearing on Many Eyes, she said, a different user uploaded the sexual harassment testimony of President Bill Clinton (see "I don’t know" and "I don’t remember" dominated the first words of his sentences.

Many Eyes offers users about a dozen templates, including maps, graphs, a scatterplot and a network diagram. One user used a bubble chart to portray the McDonald’s menu. Imagine, Deluxe Breakfast or Big Mac? The McDonald’s menu visualized as a lipid ball has the breakfast in the middle, big and fat, while the Big Mac is out on the edge, a veritable carrot stick by comparison (see Another visualized Shakespeare’s favorite words in a tag cloud, and still another used a treemap to show the visualization conference’s attendees by country.

Data presented this way starts conversations, say proponents. Users respond with opposing data sets, and observers post responses in the Many Eyes forums or talk among themselves about what they see. As a table of numbers, the same data would leave most people cold.

Other sites include Swivel, Data360, and Dabble DB. Gapminder, not open to users’ data, uses animation to show per capita relationships among all countries over time between income, internet users, carbon dioxide emissions, life expectancy and other metrics. Test your theories.

We Feel Fine is "an exploration of human emotions in six movements" and an interesting experiment in analyzing text. It’s made of phrases harvested from blogs around the world in sentences that begin with "I feel" or "I am feeling." The system records the whole sentence and the writer’s age, gender, location, and the area’s weather. The site’s mission page claims to record "15,000 to 20,000 new feelings per day." As of early November, it had collected 9,204,703 feelings.

Randomly bouncing particles represent one feeling expressed by one individual, and each particle’s color, size, shape, and opacity represents the feeling. Click a particle to read the sentence. All together, it represents "pictures of human emotion," which can be filtered by country, state, age, gender, and other factors.

It says Las Vegas is the sexiest city, Denver is the angriest, Roanoke, Virginia is the greatest, and Evansville, Indiana the most loved. The happiest state is Hawaii at 32.3 percent above average. In early November, most bloggers around the world felt "ready."

"Getting ready" might have been the headline for the major part of the conference: presentation of research papers. Though obviously not intended for commercial shrink-wrapped distribution, each paper offered new insight and a potential building block for some future product.

The work named "best paper" charted the movements of 80 office workers over one year within a 32,292-square-foot office space, and made browsing the data flexible, intuitive, and fast. "Visualizing the History of Living Spaces" described how the Mitsubishi Electric Research Laboratories team from Cambridge, Mass. figured out how to, for example, follow one individual at a given time of day across multiple spaces, quickly untangle that person’s path from the paths of others, and understand it all at a glance.

A team at University of North Carolina at Charlotte improved the way users can understand urban areas. "Legible Cities: Focus-Dependent Multi-Resolution Visualization of Urban Relationships" by Remco Chang and four others describes how they can give a one-glance grasp of space and data. With that, a user can see demographic similarities among neighborhoods or test the effects of a new school on a neighborhood’s social structure.

"NodeTrix: A Hybrid Visualization of Social Networks" by Nathalie Henry and other team members in France and Canada visualized social networks in new depth. One case study, for example, showed how two research teams collaborated differently. In the Massachusetts-based team, one dominant person was the hub, the only link that other team members had to each other. Two other teams, in Berkeley and Palo Alto, Calif. showed researchers in strong contact with all other team members.

Still other papers were interesting but not yet better than plain numbers. One Istanbul, Turkey study mapped the demographics and socio-economics statistics of different countries to hair styles.

Curiously, when asked how some of their arcane research might be applied, some grad students stumbled. One student from Germany proved he was perfectly fluent in English until I asked him how his 3D modeling research might be applied. He didn’t understand. I tried the question several other ways and finally gave up.

"Crossing the chasm" was on some minds. Visualization pioneer Ben Schneiderman from University of Maryland echoed Geoffrey Moore’s book by that name and urged infovis researchers and others to create a beachhead in the business world.

In dashboard expert Stephen Few’s capstone address, he urged researchers to "zoom out" to have a wider view of the problems to be solved. He said, "Show the people in business what you have."

Why doesn’t business appreciate information visualization more? Visualization pioneer Schneiderman said hard numbers are confused in business with hard science. He evoked Geoffrey Moore’s book Crossing the Chasm to urge proponents to create a "beachhead" in the business world.

Appropriately, InfoVIS 2007’s closing reception occurred at the nearby California State Railroad Museum. It’s dedicated to a technology that, despite the doubts of some early 19th century business people, ultimately proved more conducive to commerce than the horse.

Must Read Articles