I'll Show You My Data If You Show Me Yours

Will the web 2.0-ness never end? Now we're visualizing shared data sets, with two new projects just launched that encourage users to upload their data sets and map them against each other. Why would you want to do that? Take a look at the map below, which a user on one of the services, Swivel, created to show the relationship between global GDP and yearly average global temperatures. Interesting, no?

Data visualization may sound a bit complicated and off-putting, but it's all about making information sets easier to grasp. Instead of looking at a bunch of tables and numbers, you look at a picture which depicts those tables and numbers. Some simple well-known examples of this are the beloved pie chart, the bar chart, and the x/y graph, although more intricate data visualization can involve graphics, colors, maps, and other design elements. Also sometimes known as "information design" by the dedicated followers of visualization kingpin Edward Tufte (of which we at janethaven.com are one, incidentally), the display of data in quick-to-understand graphics is a skill worth exploring. Better yet, good information design allows you to set apparently unrelated data sets against one another to tease out relationships that are not necessarily obvious in a table of figures set side-by-side.

Data visualization tools have been on the web for some time now. From govcom.org's Issue Crawler to Hans Rosling's GapMinder to Google Lab's new Trends visualization project (here's one on searches on Repblicans/Democrats) to Data360, which has been around for about a year, there are lots of tools out there to let you look at data in graphical format.

Love of visualized data sets, however, is clearly a growth business, if the launch of two web 2.0-style data sharing-and- visualization services, Swivel and Many Eyes is any indicator. Where Flickr encourages you to share your photos, and youtube your videos, Swivel and Many Eyes both want you to share your data sets, and then visualize them. Swivel encourages you to mash up various data sets, while Many Eyes lets you work with one data set at a time, but with more options for visualization tools than Swivel currently offers. Both of them are very recent launches -- Swivel in early December 2006, and Many Eyes (a project of IBM's Collaborative User Experience research group) in January 2007.

Both projects also emphasize the social value of sharing data. Many Eyes explains:

Many Eyes is a bet on the power of human visual intelligence to find patterns. Our goal is to "democratize" visualization and to enable a new social kind of data analysis. Jump right to our visualizations now, take a tour, or read on for a leisurely explanation of the project.

All of us in CUE's Visual Communication Lab are passionate about the potential of data visualization to spark insight. It is that magical moment we live for: an unwieldy, unyielding data set is transformed into an image on the screen, and suddenly the user can perceive an unexpected pattern. As visualization designers we have witnessed and experienced many of those wondrous sparks. But in recent years, we have become acutely aware that the visualizations and the sparks they generate, take on new value in a social setting. Visualization is a catalyst for discussion and collective insight about data.

Great. Swivel is even hoping to make some money off their service, by allowing public data accounts to be free and private data accounts to be run for a fee. Both services also encourage community and data-sharing across platforms: you can blog your visualizations with copy-and-paste HTML, and Swivel is even more hooked into the web2.0-ness of it all with community features and automatic Google and Wikipedia search links. For more on the similarities/differences between the services, see the post on Tim O'Reilly's blog from a couple of weeks back.

The question all this activity around social visualization of data sets raises for me is whether people are seeing the information around them in more structured terms. To put it another way, I wonder if more people will come to these tools without their own data sets, play around with what's up there already, and go back to their own work with a new eye for what they might be able to extract usefully from the babble of infomation that surrounds us all -- or will these types of sites only appeal to people who are already data geeks, and who already see the world in terms of what data they can scrape, create or download from publically available sources.

This is an important question in my work as one of the problems we've been thinking about at the Civil Society Communications project is how to get non-profit organizations who often collect large amounts of data for advocacy purposes to think about visualizing that information rather than only collating it into a written report or a set of flat tables. The written report is important to establish a baseline set of facts and to look at trends in detail, but the information visualization piece, which is almost entirely missing from the work of most advocacy groups, particularly those working in the global south, can quickly catch the eye of new supportors and decision-makers alike. These types of organizations may not even see the information collection they are doing as generating data sets, and depending on how they go about it, they may miss that opportunity...if you think you are collecting information only for a written report, you might collect it, store it, and categorize it quite differently than if you are thinking of using it to tell a visual story.

So my hope, when I look at these types of tools that "democratize visualization", is that they will not only fulfill their stated mission, but also help with education and inspiration among those who may not yet find themselves toe-tappingly excited when someone mentions "data sets" and "visualization" in the same breath.

Disclosure: My employer, the Open Society Institute, a private grant-making foundation, provided financial support to the development of the Issue Crawler software mentioned above.