When visualizing data, beware of gratuitous rendering
- By Joab Jackson
- Aug 24, 2009
"When you are dealing with massive quantities of data, you want to make sure you are taking a visualization technique that is appropriate for your analysis task," said Andy Hoskinson, vice president of technology strategy and consulting at Unisys. When he was a field artillery officer in the Army, Hoskinson worked with geographic information systems on fire direction drills.
All too often, Hoskinson has seen the gratuitous use of graphs, pie charts or other techniques that, although visually appealing, did little to advance the agency's message.
"Instead of having visualization for the sake of having the visualization, figure out the message you are trying to convey and pick the most appropriate visualization type," he said.
Recently, Web 2.0 has brought a slew of new ways of depicting data — some better than others. Popular approaches include tag clouds, in which key words are placed on a backdrop, with the heaviness of their type signifying their usage within the dataset. Tree maps are another new technique in which topics are arranged in a simple branch-like fashion, showing their relationships to one another.
But it's important to make sure the form matches the content. "A lot of these techniques are unproved, so it is unclear how effective they are in terms in conveying messages about a large dataset to general users," Hoskinson said.
Joab Jackson is the senior technology editor for Government Computer News.