At TED2010 this week, Live Labs Director Gary Flake presents an experiment in data visualization that allows people to make more sense of the growing amount of information that surrounds them.
A Microsoft technical fellow and director of Live Labs, Gary Flake has made a career out of building technologies that glean insight from information.
Today at the TED2010 conference in Long Beach, Flake will show off Live Labs’ latest development — a data visualization technology called Pivot, designed to help people make better use of digital information.
“We’re at a really interesting point in the history of the Internet,” says Flake. “Pivot was not possible to build five years ago. But it is possible to build today.”
Pivot is an experimental technology that allows people to visualize data and then sort, organize and categorize it dynamically. The result is that correlations, exceptions and trends become immediately apparent in ways they can’t when information is stuck in rows and columns.
The program is designed to provide a much more natural way for humans to digest large palettes of information without losing their way — an idea that anyone who has analyzed giant spreadsheets may welcome. And while it has something in common with spreadsheets and many other technologies, Flake says Pivot is so new and different that it’s difficult to even picture what it does without seeing it in action.
“With Pivot you can interact with data in a way that is powerful, informative and fun,” Flake says.
Pivot combines related data — anything from pictures, videos and maps to batting averages and financials — into large collections that can then be manipulated, sorted, filtered and examined visually. In this way, the data itself can help shape and inform the way it is presented.
Perhaps Pivot’s most compelling feature is its ability to smoothly and quickly arrange collections according to common characteristics and then zoom in for a closer look, by either clicking on a particular item or filtering the collection to get a subset of information.
“With Pivot you can swim through the data, taking little twists and turns,” Flake says. “If you are looking at all the information at once, the proverbial forest, you can click on any one item or filter and smoothly zoom into the trees without any interruption.”
The example right of a Pivot view makes it clear how pitchers outnumber other players on Major League Baseball teams. Subsequent views can display salary, performance against salary, and other information on the fly.
Flake says that Live Labs’ research with users has found that the continuity and smoothness Pivot provides in surfing through data is important to help users understand what they’re looking at, and how they got there.
“We found that if you make it a sudden transition, people lose their way,” he says. “But if you make it very smooth and continuous, people have a mental model of how they got to where they are.”
Because Pivot works with almost any kind of data, its potential uses are as varied as the types of information available today — in other words, practically infinite. A Pivot collection designed to help study the history of movies, for example, could sort by male or female leads, and then sort again to find their most frequently occurring co-star. The user could then “pivot” the information again, to shape the display of the movies themselves, perhaps by decade.
Realizing that the horizon for Pivot’s possibilities was far beyond what one research group could hope to imagine, Live Labs released the technology on a limited basis last fall at the 2009 Professional Developers Conference. After only a couple of months, says Flake, the examples started rolling in.
“Just as we anticipated,” Flake says, “people are using it in ways that we never anticipated.”
Live Labs has made Pivot downloadable to people who want to give it a try and think about new scenarios it enables.