We could have come across a very famous saying “a picture is worth a thousand words” so many times in our lives. Let us dive into our memory when we are at the age of 1 - 8 years old. Do we remember any details of the conversations we had? Not most of us do. Do we remember a picture and heard family members narrating stories around it and remember those stories or the picture itself. Yes, most of us do. This is because we remember or associate information better with pictures or visualizations. A lot of scientific experminets were conducted on this subject and lot of companies emphasize on visual presentations to get few key messages to employees or public.

Let us dive into one of the researche papers - Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods. The paper defines the graphical perception as the vissual decoding of the information encoded on graphs and includes both theory and experimentation. The paper looked into list of perpetuary tasks that people perform to extract the quantitative information from the graphs and then look into accuracy of extracting the information. The paper showed us the basic foundations of how humans perceive the elements of graphs. The below figure illustrates 10 elementary perceptual tasks that people use to extract quantitative information from graphs. As we can see from the figure, humans have a better judgment on dot position than length, direction, angle, area, curvature and volume. This “elementary perceptual tasks” clearly showed how we visually-mentally process graph elements.

knitr::include_graphics("C:/Users/Gautam Reddi/Dropbox/Harrisburg/Term 2/Data Visualization/Perception Essay/Figure1.jpg", dpi = NA)

Let us now look into the Heer and Bostock paper about assessing the viability of crowdsourcing for evaluating the large design space of data visualization. They assessed the viability of Amazon’s Mechanical Turk as a platform for graphical perception experiments. They firstly replicate previous graphical perception experiments and their crowdsourced results are a good match. Although the variation of the results are increased, the scalability of the platform can offset it. Their research shows new insights for visualization design. It also proves that crowdsourced perception experiments are viable.

I have taken the research paper “Introduction of Human Perception in Visualization” as oen of my two papers as I believe visualizations shopuld be created to help humnans synthesize data and take decisions based on it easily. The paper explains three activities that are supported through visualization tools - Exploratory analysis, Confirmatory analysis and Presentation.All these activities together help to explore given data, determine evidence to accept or reject data formed through previous analysis and present graphical display of the structure, behavior, relationship, and other essential characteristics of the data involved. The paper also talks about why visualization helps in analyzing the data faster as Vision is the first component of the human sensorial system and it is the one that is acquired fastest by the human brain. The paper also dives deep into the shape and space perception by humans.

knitr::include_graphics("C:/Users/Gautam Reddi/Dropbox/Harrisburg/Term 2/Data Visualization/Perception Essay/Figure2.jpg", dpi = NA)

In the above visualization reference model, we can identify three phases of the visualization process: data transformation(s) (pre-processing), visual mapping and view transformation (rendering).

In the second paper, I have taken “an evaluation of non-expert viewers’ data literacy” written by Zubiaga, Arkaitz and MacNamee, Brian as the ability to interpret the outputs from data analysis is important when we present data. Data analysis can be interpreted differently by different types of audiences, based on their expertise. That is why perception study is more important. This will help us to build data visualization that can make similar data interpretation in all the audiences. One of the experinents in the research shows most of the users achieved 97 percent accuracy with histograms, as compared to other charts. In another experiment the research indicates that histograms can be more suitable than density trees for visualization of data.

knitr::include_graphics("C:/Users/Gautam Reddi/Dropbox/Harrisburg/Term 2/Data Visualization/Perception Essay/Figure3.jpg", dpi = NA)

All these different researches and experiments show importance of graphical perception in data visualization and provides us understanding of the types of visualizations that we can use.

References

  1. W. S. Cleveland and R. McGill. Graphical perception: Theory,experimentation, and application to the development of graphical methods. J. Am. Statistical Assoc., 79:531-554, 1984.
  2. Crowdsourcing Graphical Perception: Using Mechanical Turk to Assess Visualization Design Jeffrey Heer, Michael Bostock ACM Human Factors in Computing Systems (CHI), 203-212, 2010
  3. Alexandre D. and Tavares J. (2010, 27 May) Introduction of Human Perception in Visualization
  4. Zubiaga, Arkaitz and MacNamee, Brian. (2016) Graphical perception of value distributions : an evaluation of non-expert viewers’ data literacy. Journal of Community Informatics, 12 (3). ISSN 1712-4441