Jane's Colloquium 2

The Thrill of Discovery (October 4th)
Ben Shneiderman

We are inundated with data. While we quickly work to develop new tools for data capturing, we are still struggling to process, understand, and partially digest mountains of data that have been collected for decades. The major problem lies not in having too much data, but in not having easy ways to visualize and interpret millions of data simultaneously. And even if we could, how do we effectively break down the data to view the influence of various contributing factors?

Ben Shneiderman, presented various methods he, his student’s, and his colleagues have developed to facilitate this overwhelmingly common problem. By creating user-controlled interfaces which first provide a high-level overview and then allow users to telescope in on areas of interest, while providing optional filtering of data characteristics and detailed information on demand, they have created a highly flexible and scalable user interface. He provided a successful example of this implemented on www.spotfire.com. Additionally Dr. Scheiderman showed examples of a treemap data visualization structure developed at the University of Maryland that has been widely successful and used on such sites as www.smartmoney.com/marketmap and Pete’s coffee website.

Dr. Shneiderman mentioned three necessary characteristics for a successful user interface: rapid, incremental, and reversible. By rapid he is referring to the ability to query data quickly and change functions without delay. Incremental refers to starting with a broad data view and providing ways to scale down to view a smaller, more detailed data sets, as well as zoom back out. And last but not least, it is important to always provide a way to undo or reverse actions.

Last modified 25 November 2007 at 2:26 pm by jane