‘A thought, belief, or opinion, often held by many people and based on appearances’ - this textbook definition of perception from the Cambridge Dictionary has been long since used by drivers of strong influencer relationships to create positive mental impressions, be it in marketing to attract attention of customers, finance to attract attention of investors, or even academics to attract attention of grant providers. The consistent pattern in intent here is ‘attracting attention’. Perception, like communication, can be viewed as cyclical, involving initiation and feedback. The key here is shaping it such that receptors can recognize, organize and interpret intellectual, sensory and emotional data in a logical or meaningful fashion. Effective communicators often use sensory perception to generate the required stimuli at the other ends of communication channels - through the use of sight, hearing, taste, smell and touch. While activating any of these senses would be impactful, in this essay, we will be focusing our attention on visual tools and analyze the role of perception in visualization.
Cleveland and McGill attempted to scientifically formalize the study of visualization- suggesting that there is a need for scientific approach to the subject, and for the theory of graphical methods. They introduced a theory that certain types of graphical representations can have a significantly higher impact on viewers’ judgement of quantitative information over other forms. For this, they used ordered sequencing to redesign old graphical forms to offer alternatives graphical forms and assess whether specific perceptual tasks are critical factors in determining the performance of a graph. They found that as the distance between two values being judged increases along an axis, the accuracy of viewers’ judgement decreases, and that the position judgements were more accurate than length judgement. They also found that dot charts are likely to be perceived more accurately compared to pie charts or even bar charts, and that bar charts are preferable to pie charts or a divided bar chart.
Heer and Bostock reviewed the relevance of crowdsourcing for evaluating visual designs and charts, and found it particularly attractive as it is an inexpensive, often instant and an effective method to obtain useful insights for visualization design. They did, however, recognize the need to validate the credibility of participants, as crowdsourcing brings about issues related to as ecological validity, subject motivation and expertise, and lack control over many experimental conditions. They tested the feasibility of using Amazon’s Mechanical Turk to evaluate visualizations and then used these methods to gain new insights into visualization design. They found that crowdsourcing of graphical perception studies can be viable. The increased variation of results was compensated by the platform’s scalability, as many more subjects could participate for the same cost.
Heer, Kong and Agarwala in their 2009 article - ‘Sizing the horizon’ investigated effectiveness of different techniques for visualizing time series data. They compared line charts with horizon graphs-a space-efficient time series visualization technique-across a range of chart sizes, measuring the speed and accuracy of subjects’ estimates of value differences between charts, and found we find optimal positions in the speed-accuracy tradeoff curve at which viewers performed quickly without attendant drops in accuracy. They found that mirroring a chart, i.e. flipping the negative values around zero did not affect estimation time or accuracy. As mirroring cuts the size of the chart in half without any observed downside, they advocated its use when space constraints warrant, so long as the viewer knows how to interpret the chart. They also noted that layered bands are beneficial as chart size decreases, even though generally, at constant chart heights, layered bands increase estimation time and error. In addition, they observed that that smaller sizes led to faster estimations, and that for each chart type there is at least one size that minimized estimation time while preserving accuracy.
In their 2008 paper ‘Color Design for Illustrative Visualization’, Wang, Giesen, McDonnell, Zolliker and Mueller presented a knowledge-based system that captures established color design rules into a comprehensive interactive framework, aimed to aid users in the selection of colors for scene objects, while incorporating individual preferences, importance functions, and overall scene composition. The framework also offered new knowledge and solutions for mixing, ordering and choosing colors to present semi - transparent layers and surfaces. The framework was created keeping in mind prominent guidelines for visual color designs, well-known facts in the industry and insights from human visual perception. In addition to establishing color harmony and compliments, other rules pertained to vividness of colors which stand out but also can be overwhelming, Foreground-background separation, differences in hue, saturation and lightness, etc. After selecting the hues from a color wheel, the system runs a conversion algorithm for vividness and lightness for class and class component highlighting and intra-class contrast. By formalizing a set of color design rules, the system is expected to achieve more task-effective and aesthetic color designs with ease, and eliminate the trial and error process that comes with picking the ‘right’ colors from a set of millions.
Significant research efforts are in play to understand the role of perception in visualization. Considering that the ultimate goal of any graphical presentation is effective communication, understanding the human brain and how it processes information presented to it is key in resonating with the end user of the information. It is, however, pertinent to note that people function differently, which brings about considerable judgement and variability in the study. Nonetheless, a set of rules that originate from extensive research would provide a well-tested channel with a higher likelihood of success in connecting with the target audience.