Visualization as Technology, Economics, and Art
It’s a simple recipe… with lots of ingredients.
On my first day of the Data Visualization (INST 760) course at University of Maryland, we were asked to introduce ourselves and explain why we were interested in the subject. For about 80% of the answers, I felt like we were all giving a variation of the same answer: “I suck at producing visually appealing and easy to understand visualizations, and I want to learn how to do it better.”
Without any prior experience in visualization, if not for the wonderful graphics I have occasionally created with Excel and the like, I took a dive into a paper by Jarke J. Van Wijk that was so simply titled as “Views on Visualization.” This is an attempt at unpacking what I read.
There is no question that visualization is the simplest way for your regular Joe to obtain insight from data through perception of patterns and other interesting features such as colors, shapes, groupings, etc. However, by considering the context in which a certain visualization is used, we can assess what its actual value is. Given the multiple views on the growing field of visualization, we ought to ask ourselves what the status is, and what the possible implications and directions are.
The economic model that van Wijk puts forth is one I had not ever thought about. It suggests that the knowledge derived from a certain visualization depends on the graphic produced, the individual’s prior knowledge, and their perceptual and cognitive abilities. A very good example given is that a map showing a province of the Netherlands provides more information to a person from the United States more than it does to a Dutch person. The author takes this example further to add elements of perceptual and cognitive abilities by showcasing that a person who is visually impaired will be less effective at reading the map correctly. The economic model suggests that for a visualization method to be considered valuable, it must be used by many people who use it frequently and obtain highly valuable knowledge from it without spending time and money on hardware, software, and effort.
We all want visualization to be like that magical wand that gives us all the information we need, but like any other process, there are two significant problems that professionals must surmount: (1) visualizations are subjective, and (2) they can be wrong and/or misleading. For the former, we have already established that prior knowledge, as well as perceptual and cognitive abilities of the user, has an impact. Therefore, the value of a certain visualization is almost always going to be subjective, depending on who is performing it. The latter tells us that we need expert input on visualizations so that they convey sound messages.
Van Wijk also puts forth an interaction viewpoint, stipulating that good defaults and automatic ways to set visualization parameters should lead the user to gain as much knowledge as possible before they perform any other customization. In the main uses of visualization — exploration and presentation — we should aim at producing the most insight for our users without a significant expenditure on their resources.
I once took an advanced usability testing course where a fellow classmate produced one of the most elaborate user journey maps I had ever seen. I honestly left that class session feeling like I did not have the artistic range, nor the innovative thinking, to ever pull off such aesthetically pleasing and highly informative graphics. I made it my personal goal to learn visualization so that I could at least tickle my creative senses into thinking graphically. Van Wijk’s paper teaches me that it takes effort to think about prototypes, and even more effort to decide on which product is successful for your purposes.
In the end, what I take from van Wijk’s paper is the understanding that the various views on visualizations that he presents are all related, and that as the field grows, we should continue to aim at accomplishing effectiveness, efficiency, aesthetics, and generic laws with predictive power.
If I were to go back to that first day of class and get to answer my professor’s question again, I would say that I want to learn how to understand the audience that will use my visualizations, understand which viewpoint they are looking at my work through, and finally how to best tailor my work to their needs. This paper helped me articulate these thoughts:
- Jarke J. van Wijk. Views on Visualization. IEEE Transactions on Visualization and Computer Graphics, 12(4):421–433, 2006.
