3D visualizations are generally a bad thing. They present a distorted representation of data, and they are hard to perceive correctly. Most of this stems from the fact that they are in a virtual 3D space, but are projected onto a 2D screen. But what if there was a way to make them actual tangible objects in the real world? Would they still suffer from the same issues?
This is what Yvonne Jansen, Pierre Dragicevic, and Jean-Daniel Fekete set out to discover with their research into physical visualizations (to be published at CHI 2013). Their study uses both digital and physical visualizations, identical in every regard other than existing in virtual vs. physical space.
They tested various interaction techniques on each, including combinations of the two, where the physical visualization acted as input for manipulation of the digital version. Their results suggest that physical visualizations are reasonable ways to communicate data, and they outperform their virtual counterparts.
One reason for this may be skipping the projection to 2D space, but there are other possible reasons. Direct manipulation is a powerful way to interact with something, and it can help to reduce cognitive load by removing mental abstraction and memory tasks.
The practicality of creating physical visualizations is definitely questionable. Yvonne’s team used digital fabrication to create their models, but this is not a technique that is appropriate for all circumstances. Some visualization techniques would require floating objects. And physical things cannot be shared in massive quantities like digital visualizations. It takes a lot of resources to create them, and they cannot always be easily manipulated to represent different data.
There definitely are applications where physical methods are a good technique, and Yvonne and her team have been building a list of physical visualizations that exist. Data that is fairly static, or specific to a location is a great candidate for physical visualizations. Data with complex spatial relationships is also great. Data driven architecture or sculpture is also an interesting place for physical visualization. And as robotic output devices become cheaper to create, it may be appropriate to have “tangible displays” that change form based on data.
Perhaps the most important part of this work, though, is what we learn from it about our brains. Knowing how different input methods are used in our brain is important to understanding how we work, and improving our lives.