Graphical/Schematic visualization is crucial part in statistical data analysis. It is widely used in several fields such as supply chain, research, business consulting. As a business consultant, I use visualization everyday to convert data to information and create story lines for clients. We have been using graphs over last two centuries to represent and visualize data. Although there is a lot of scope for improvement and most of the illustrations are not representative of true data. There are several ways to represent numerical data apart from graphs. Graphical representation can be really useful in extracting patterns and predicting behaviors that may not be revealed by just performing analysis on raw data in tabular or list form.

Perception of visualization has been discussed and studied for a long period of time in human’s development of data processing and display. The main purpose of data visualization is to assist in better decision making. Charts are used in financial and forecasting industries. Converting data to information is also known as drawing insights from data.

Tremendous amount of research has been conducted on visualization parameters such as length, width, area, color, position and shape as well. Lot of experiments have been done to understand the graphical methods and our perception on graphical representation, one such experiment was done by Cleveland and McGill. Perception theory was tested by them. Graphical perception deals with graphics and how humans interact with them psychologically and emotionally. The relation between human perception and graphic visualization can be termed as graphical perception. A graph can be developed by using either qualitative or quantitative data or both. The same data is decoded when the person sees the graph visually and perceives the graph based on their thought process and interpretation of results. Human perception plays a crucial role for quantifying graphical perception. As a result of these experiments, it was concluded that judgement based on length is much better than a judgement based on area. It was also found that judgement based on area are much better than that of volume based judgement.

Several research studies have been conducted and several are in progress based around the graphical perception theory of Cleveland and McGill. One such graphical perception research is Crowdsourcing. This research uses a platform created by Amazon known as the Mechanical Turk also known as MTurk. A pool of users is ready to work on any tasks that are posted by the researchers. A reward is presented to users for completing the task as a perk for the collaborative effort. Major goal of the crowdsourcing is to simulate perception on historical experiments. Mechanical Turk research demonstrates that it is feasible to conduct graphical perception experiments by utilizing crowdsourcing as an alternative and a much better option. Some of the key things that came as a by-product of the crowdsourcing experiment was also shed some light on aspects on perception towards rectangular area visualization.

Source: Wikipedia: https://en.wikipedia.org/wiki/Ebbinghaus_illusion#/media/File:Mond-vergleich.svg

When it comes to perception of circles it is very much needed for a cartographic environment type of work. Patricia Gilmartin researched this topic in her paper titled ‘Influence of map Context on Circle Perception’. Several parameters have been identified to see the effects of circles and dots on maps in geomapping. For example, the difference in the size of a target circle and the circles around it causes a perception issue. The target circle looks smaller when a group of adjacent circles around it are larger. This is in contrast to when you have a target circle that is surrounded with smaller circles around it.

Data literacy is one more effective concept of Data visualization that is key when it comes to understanding the results of a complex data analysis. Perception and data literacy go hand in hand. When we consider perception parameter when it comes to data literacy, it helps to address a wide variety of audience when it comes to interpreting data visualization models. Zubiaga and Namee have used Crowdsourcing platform to conduct an experiment that demonstrates that most of the users achieve 97 percent accuracy when it comes to histograms, as compared to any other kind of graphical charts. In a second experiment, their results demonstrate that histograms can be more effective than density trees for visualizing complex data.

Conclusion: Graphical perception is key to understanding the deeper concepts of data visualization. Data visualization needs to be tailored to the audience keeping in mind their perception of the information we are trying to convey to them.

References

W. S. Cleveland and R. McGill. Graphical perception. J. Am. Statistical Assoc., 79:531-554, 1984.

Crowdsourcing Graphical Perception: Using Mechanical Turk to Assess Visualization Design Jeffrey Heer, Michael Bostock (CHI), 203-212, 2010

Accessed from the internet on 06/24/2017: Source: Wikipedia: https://en.wikipedia.org/wiki/Ebbinghaus_illusion#/media/File:Mond-vergleich.svg