IBM's Many Eyes project seeks stats freaks

The Many Eyes Web site, in public alpha since January, enables people to upload data sets from common delineated files, such as an Excel file, then go through a few clicks to make it visual and available to all.

There are 15 different types of data charts, called visualizations, that can further be manipulated by viewers to search and parse for different views and more specific data subsets. Users can upload their own data set, and also choose others' data sets to visualize.

The idea behind the site is that people who have a common interest in certain types of data can come together on the Internet as they do with music, movies and video. The group hopes that it will be a place for researchers to make contact, too.

"We think people will use it to build communities around data," said Matt McKeon, a developer for Many Eyes.

Toward that end, the IBM research team is adding more features for social-networking and Web 2.0 ease of use.

An option exists for pulling the visualizations and tools into blogs, but users have requested bigger versions and live updates to incorporate into their Web sites. The Many Eyes team plans to add a ratings system for the best, as well as static tag categories and more browsing options. (Currently, all the data sets and visualizations are tagged by their creators.) Many Eyes has a place for comments, but the team plans to add categorized forums where people with common interests can discuss data.

A number of these and other social-networking components should be in place by the summer, according to McKeon.

IBM seeded it initially with about 20 data sets, but now there are over 2,000 data sets from users, according to McKeon. McKeon originally worked on a similar data visualization project for the U.S. government in Iraq to help soldiers evaluate combat data.

"One really surprising thing is that the Christian bloggers found us. They have been using it for dissecting the Bible...And then there's the book people, a lot of stuff uploaded from Project Gutenberg. We thought sports statistics would be the big thing. There have been some baseball and hockey stuff uploaded. But yeah, when the sports statistic freaks find us, that will be nuts...and very cool," said McKeon.

Many Eyes has data sets from the U.S. Census Bureau, the U.S. Office of Management and Budget, the Department of Education and even Linden Lab. One of the charts shows which countries Second Life residents reside in, in their first life.

Irene Greif, IBM fellow and director of the collaborative user experience at IBM Research, said at an IBM press event on Wednesday that Many Eyes seeks to "harness the collective intelligence," the tag line on the Many Eyes home page.

"If people can collectively share and learn, then will issues that are not normally understood, be understood by more people? That is the intention of Many Eyes, to give people a comfort level with sharing data," said Greif.

When it comes to legal issues of private data, McKeon said that the group complies quickly with any take-down requests. So far, there have only been a few, from early users who did not realize their data would be public. Since then, the team has added more warning signs to its pages.

The team is also working on the ability to host data that could then be kept private or by invite-only. Organizations are interested in using Many Eyes as a hosted service to disseminate information to the public. IBM is also in talks with companies interested in a private version for internal data access, said McKeon.