Perception Essay: The role of perception in visualization

Visualization in the form of graphical methods is needed for data analysis and data presentation and which has been in use nearly 200 years. This method helps human in revealing data patterns and structures which may not be revealed by just looking at the numbers, tables or statistical measures. The scientific reason or foundation behind this is how we human process the information available in graphical form which is nothing but graphical perception.

Cleveland and McGill have done some experiments to test the human understanding of graphical methods in Graphical Perception Theory, Experimentation, and Application to the Development of Graphical Methods paper [1]. They explained that graphical perception plays a significant role in visualization as it is purely based on decoding visual information which is encoded in graph. They found a set of primary perceptual tasks to extract one of the important quantitative information from graphs and tested by experimentations to prove that specific perceptual tasks are critical factors in interpreting visualizations, patterns and behaviors. They discovered hypothesized ordering of perceptual task based on different information from various sources and how accurately human perform to provide guidelines in designing graphs. From these experiments, they reported that position judgements are better than angle and length judgments. This communicates the importance of relation between visualization and understanding process of graph in human-graph interface. They suggested that because of the impact of perception in decoding graph, there is a need to create more personalized and better graph forms than old designs to lead human judgments close to actual encoded meaning.

Extending the work of Cleveland and McGill, Heer and Bostock[2] introduced the new way of evaluating visualization designs using low cost, diversified and scalable crowdsourcing approach as visualization design space is large which needs wide range of participants. Their studies provided results similar to the previous results done by Cleveland and McGill as well as showed new insights for a link between visualization designs and perception. The crowdsourcing platform which imitates real world situation where pool of users work on task and get paid is proved to be feasible for graphical perception experiments. The new insights include design parameters such as luminance contrast, aspect ratio, rectangular area judgments, optimal chart heights and gridlines which proved to have influence on perception in this paper.

To enhances the effect of designing effective visualization on human visual systems, Christopher and James[3] in Attention and Visual Memory in Visualization and Computer Graphics paper explains about previous and current theories of visual memory and visual attention such as feature integration, textons, similarity, guided search, Ensemble Coding, and Boolean maps. They discussed the preattentive processing which is occurring within the visual system and human ability to remember details and deploy attention to significant details within the data have major influences on visualization. In this paper, author used different visual features and showed how certain visual features favored by visual system over others. Visualization and results from psychophysics goes hand in hand which help to visualize important clues in the data.

Adil, Niklas and Benjamin[4] identified six cognitive exploration stages which incorporate perception as an assessment activity in visual data exploration.

In above figure 1[4], assessment activities include the measurement of courses of actions taken, data visualizations through perception, the changes in the interface, and include reasoning on whether the analytical goals have been achieved for available data. Authors also discussed barriers to the framework and used this framework to understand the effect of cognitive stages on established graphic design guides. Whereas Bahdor et.al.[5] studied about the effectiveness of interacting with graphical encoding (e.g., changing distances between points in a scatterplot) on perception. They compared their study with previous studies and found that participants achieved more confidence when decoding the graph with the help of interactions. In their study, they used 12 perceptual tasks (Figure 4[5]) which have higher accuracy for interaction purpose. Their study observed a significant difference between length and angle in terms of accuracy when compared to prior studies. One of the interesting point mentioned in this paper is input from perception continually impacts interaction. Hence, the perception of the encoding might be directly connected to the performance of interaction with an encoding.

From all these studies and papers, we can conclude that to have effective and efficient visualization design, it is very important to understand perception. By understanding perception role in visualization will help to make similar and accurate data interpretations and close to actual encoding by different types of populations. Better visualization will accurately encode all qualitative and quantitative information with the help of graphical perception theory which will make it easy and simple to decode.

References:

[1] Cleveland, William S., and Robert McGill. “Graphical perception: Theory, experimentation, and application to the development of graphical methods.” Journal of the American statistical association 79.387 (1984): 531-554.

[2] Heer, Jeffrey, and Michael Bostock. “Crowdsourcing graphical perception: using mechanical turk to assess visualization design.” Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, 2010.

[3] Healey, Christopher, and James Enns. “Attention and visual memory in visualization and computer graphics.” IEEE transactions on visualization and computer graphics 18.7 (2012): 1170-1188.

[4] Yalçin, M. Adil, Niklas Elmqvist, and Benjamin B. Bederson. “Cognitive Stages in Visual Data Exploration.” Proceedings of the BELIV Workshop: Beyond Time and Errors-Novel Evaluation Methods for Visualization. 2016.

[5] Saket, Bahador, et al. “Evaluating Interactive Graphical Encodings for Data Visualization.” IEEE Transactions on Visualization and Computer Graphics (2017).