September 20, 2012
A friend of mine recently mentioned that she’d attended a workshop on visual data. Visual data? I thought she meant infographics and in some small way she did. But she went on to explain the larger picture. By visual data she meant the art of using motion graphics, still pictures and other interactive, inventive and artistic techniques to visually represent deep data sets and complex data records, facts and figures. Intrigued, I investigated further.
I came across a number of interesting cases worth sharing. First, I stumbled on something called a wind map. Using the National Digital Forecast Database of weather information, two artists developed the map as an art project. Their web application collects real time data from NDFD and turns it into a map that updates continuously as we watch on. It shows the wind crossing the Rockies and swirling around the Gulf. This still photograph doesn’t do the concept justice, but at least offers a conceptual idea of what to expect.
Another example of visual data comes from an app built by Marcel Salathe. The app reads the HTML source code behind websites, assigns a color to each type of tag in the code, maps the way the tags nest inside each string of code and then charts a pattern depicting the data. By typing in the web address of your favorite site, the app renders the diagram before your eyes.
For my last example, I turned to fashion. I recently helped some of my Fast Horse colleagues with a project for Diet Coke. The project involved Fashion Week in New York City, which inspired my interest in this next form of visual data.
Fashion changes constantly. Each season designers select different colors to showcase, some catch on and others don’t. The MIT-based developers of this next app set up a stationary camera on the streets of Cambridge and captured the comings and goings of people as they walked by.
The software cataloged the colors of people’s clothing and created a map in the process. The computer program groups dominant colors together into color networks. The resulting graphic resembles a Jackson Pollack painting, both beautiful but more importantly representative of a vast database of colors of clothing captured on camera in Cambridge.
The developers captured the whole process in this short video.
Here is how the map rendered over time.