Stick figures enter buildings that grow a notch taller with each visit. Blocks fall from the sky into stacks. Dots cluster one by one around larger circles on the computer screen, looking like cells on a microscope’s slide.
These varied images are different attempts to visually illustrate the flow of reader activity at Digg.com, the popular news site that lets readers post links to headlines elsewhere online and rank them by voting on their popularity. Ordinarily a text-based list of stories, Digg and its Digg Labs project are increasingly experimenting with these alternative interfaces, hoping to find new ways to browse through the site’s bewilderingly rapid stream of content.
Digg Labs, which celebrates its one-year anniversary this week, has provided one of the Web’s most prominent experiments with data visualization, a young discipline that mixes graphic design and statistical analysis in attempts to represent information with pictures instead of raw data.
The Digg Labs tools, created by San Francisco firm Stamen Design with input from Digg, have drawn decidedly mixed reviews: they’re lauded by some as simple and useful, criticized by others as fascinating but confusing. But either way, the tools have kept the popular site in the vanguard of a trend that’s changing the way Net users interact with online information.
“We’re still in a phase of learning how to use these tools,” says the University of Maryland’s Human-Computer Interaction Laboratory research scientist Catherine Plaisant. “But this is finally becoming a language that people are becoming familiar with.”
Digg Labs, like most experiments with visualizing data, has been driven primarily by an attempt to cope with a stream of information abundant enough to be overwhelming.
As the site’s popularity soared over the past several years, user-submitted content began appearing and disappearing from the front page too quickly for most readers to see. Digg founder Kevin Rose commissioned Stamen to create visual ways for users to find headlines, as well as to see patterns and trends in the content flow.
The result, launched last July, has been four flash-based applications that illustrate readers’ activities. One view, called Stack, represents individual users’ votes for stories as bricks falling Tetris-like into stacks, creating a bar graph of popular headlines that grows in real time.
Another view, dubbed Swarm, offers an eerily beautiful look at how individual users cluster around popular stories. Visually, Swarm is reminiscent of cells under a microscope, with individual users attaching themselves to circles representing stories. When the same user votes for two stories, a tentacle reaches out and tugs those “cells” closer together.
As far as the analytic tools go, the company itself rates the experiments as a success. In an e-mail interview, Digg creative director Daniel Burka says that Digg Labs’ tools have influenced the company’s interface design by exposing how readers move between content areas and “swarm” around stories.
“After seeing users congregate around stories and examining their relationships, we’ve tweaked our algorithms to take [content] diversity into account when determining how popular a story really is,” Burka says. This allows a wider range of subjects to show up on the home page, for example. “Many of the lessons we’ve learned in the Labs are also influencing future feature development and the general direction of the site.”
Many readers have reacted enthusiastically. The April release of an API (application programming interface) allowing outsiders to build their own visualizations using Digg’s data has sparked dozens of independent Digg interfaces, such as Digg City, the winner of a recent contest calling for reader submissions.