The Visual Scalability Hexagon
August 27, 2014
We can see newly created data visualizations nearly every day around the world wide web. Starting from creating charts in spreadsheet applications to the creation of highly individual data visualizations, discovering new perspectives, has changed our way to see and interact with data. Because of this development, i had a question that has to be answered during the last months:
Why is the number of visual elements significantly increasing?
The first thought was the ongoing digitization of nearly each aspect of our life. This gives us new insights in our daily life that can be shown by an impressive example of Nicholas Felton. The second thought was that this hunger of generating and using more and more data has affected the creation of data visualization, too. This led to the following question:
What are the influencing factors to create a data visualization that is capable to bring a large amount of data on the screen?
It was time to open the books and to take a closer look behind the theory of data visualization. I found a relevant term about “Visual Scalability” that can be a main driver and differentiator to explain the data visualization from the past and today. Digging deeper in the literature has uncovered six influencing factors of visual scalability that can be summarized in a six-sided hexagon.
The influencing factors of human perception, visual metaphors, interactivity, visual space, data structures and algorithms, and computational infrastructure were a result by analyzing 76 papers.
Human perception: Our cognitive effort is limited to recognize a huge number of visual elements. The limitation is because of the fast recognition of form, color, movement and their underlying interpretation through a pattern. While our human perception is currently restricted the number of possible human is not. The essence of data visualization is to provide insights not only for one user but also for a group of users. This means that the impact of data visualization is depending on sharing created visualizations within a group or mass of users easily.
Visual metaphors: The number of high scalable data visualizations is increasing. If we take a look at the collected works by infosthetics, visual complexity or visualizing.org, it is visible that the number of visual elements is increasing to cover the complexity of data. This increase is changing our thinking of creating data visualizations. Past thinking was focused on a small number of visual elements to avoid an early cognitive overload. Instead, the collected works are showing that it is possible to cover the complexity of the data and to point out clear insights in the same way. This combination can be aesthetic, if existing design principles are used, too.
Interactivity: Thinking on the movie “Minority Report” reminds me that enabling the whole body beginning from using peripheral devices (i.e., mouse), multitouch, gestures to the whole body can bring the user the capability to explore the data naturally.
Visual space: “The bigger the better” is my personal choice of size to bring data visualizations on the screen. We have to consider that the limited space on the screen allows only a limited number of displayable and recognizable visual elements regarding the chosen screen. But there were a huge development on bringing a high variety of display types (i.e., mobile phone, tablet or desktop) since the first CRT display in the 50s. The rise of virtual or augmented reality and 3D printing are promising new projection areas that allows data visualization to take place in real, semi-real or virtual environments.
Data structures and algorithms: The visualization of one-, two-, three-, or multidimensional data has already established in the field of data visualization. But there are also unexplored data sources from the media. The works by Lev Manovich are showing that visualization can take up new horizons. But this is only possible, if more standardized applications are provided to develop data visualizations.
Computational infrastructure: Last but not least, a lack of performance can even frustrate the friendliest user while interacting with your most beautiful data visualization. If you care about a responsive design of your data visualization, beginning from the data processing to the frontend, the user will reward you with sharing your data visualization within his community. The challenge of visualizing a large amount of data brings the possibility to see a high number of patterns, insights or interesting data art projects but can lead to pitfalls (e.g., downtimes, blocking UI thread, long data loading) that slows down your application.
The uncovered hexagon of the six factors can be helpful when it comes to create a new data visualization and to struggle with a large amount of data. This was a short excerpt of the literature review in the last months. The term “Visual Scalability” has already raised up in 1994 that can be seen in the last figure and explored in three slides.
If you want to get a closer look into the scientific perspective of the main driver “Visual Scalability”, just take a read on the state-of-the-art paper at the conference (paper).
