SAN FRANCISCO – Like many designers, Eric Rodenbeck has had a long relationship with bar graphs and pie charts. He just thinks they are a little old school for today’s data-filled world.
Rodenbeck has experimented with animation, three-dimensional maps that show the height of buildings by color changes and a representation of how photos spread on Facebook that looks like ice crystals forming on a car window. He has even tried to characterize in a graphic how people were communicating in back channels at business conferences, with the biggest talkers at the center of a series of circles.
He is, in short, trying to rethink how data is presented.
“It doesn’t work if it’s not moving,” said Rodenbeck, the head of Stamen Design, a San Francisco studio that Google, Facebook and Microsoft have all used for help in turning fast-paced digital information into easily understood images. “It doesn’t work if you can’t touch it.”
Nowadays, devices and people are unceasingly uploading all kinds of information about the economy, locations, weather and even what sweater makes them happy. With this flood of data, some people believe that traditional ways of displaying information do not work well anymore. So there is a demand for Rodenbeck’s sort of creative thinking about the humble pie chart.
Helping people easily make sense of that fire hose of information is a challenge. Better pictures (yes, designers believe that pictures do often work better than a thousand words) can help. But many designers, particularly in the tech world, are straining for new ways to express information.
“There is no established visual language for this diversity of data sources and consumption,” said Tomoko Ichikawa, senior lecturer of design at the Illinois Institute of Technology’s design school. “People are experimenting with a whole spectrum of things right now.”
Rodenbeck’s comment about using more animation is more than an idea. Just as films are made by shooting 24 still images a second, then running them together to show movement, charts of fast-changing data reports can provide a clearer idea of the information gathered. The animated changes may be shapes on a map growing and shrinking, colors of bar charts changing or positions of lines rising or fading.
“You can’t generate a single design artifact for this much information,” he said. “You have to generate a system that shows the flow.”
Douglas van der Molen, a former Google designer who created visualizations for ClearStory Data, seeks ways for ordinary people to go on data journeys.
“You’ve got information moving at a high velocity and distributed teams with different skill sets looking at it at the same time,” he said. “Good visualizations aren’t just about representing data well, but getting people to interact with it.”
Within ClearStory’s software is an inference engine that harmonizes the different ways data are recorded and figures out a common representation for all those categories. Van der Molen’s software then picks a useful framework, like an unmarked map of the United States, where the data can be imposed. When new data is added, colors may become more intense, for example, or a particular state will pull away from the country as the information is presented.
“We spent a lot of time thinking about how fast things should shift, what we call ‘perceived latency,’ ” van der Molen said. “If things move too fast, or if the colors become too bright, it doesn’t feel like a process they are controlling.”
The more control people feel like they have over this stuff, he figures, the more likely they are to explore new features and better learn the system.
That giant tub of information is also consumed by people from diverse backgrounds. Theoretical researchers and repair workers both use General Electric’s electrical data. So how do you make it useful to both?
When the data shows something going wrong with, say, a wind turbine, “a Ph.D. looks at as a sign of a system failure and wants to think about mistakes in his calculations,” said David Cronin, director of interaction design at GE’s Big Data center in San Ramon, Calif. “A field engineer thinks about fixing something.”
“We’re looking for ways to augment human perception to help in complex decision-making,” he said.
Much like van der Molen’s ideas about data journeys, GE looks at what it calls “semantic zooms,” so different meanings appearing as data are analyzed close up. This might mean initially looking at weather, electricity consumption and maps to figure out where a power outage is likely to occur. But close in on photos at street view and it becomes clear that trees in a certain area should be trimmed so they do not fall on power lines.
But not everyone thinks the way data is presented needs to be changed. “If I need to get somewhere, I still want a map,” said Christian Chabot, co-founder and CEO of Tableau Software, a data-visualization company in Seattle. “If I want to see things in time, I need a trend line. If I want to understand the magnitude of things, I want a bar chart.”