top of page
This site was designed with the
.com
website builder. Create your website today.Start Now

Data Visualization Research

When I begin a workshop focusing on data visualization we usually start with a discussion of some foundational research that directly impacts how we think about data visualization principles. This page is not meant to provide a comprehensive overview of data visualization research, but rather to share key resources that I use when teaching.

Gestalt Theory

In the 1920s, a group of German psychologists strove to explain how we perceive information by cognitive processes such as grouping, identifying patterns, and differentiating components. Based on extensive observations, their work, called Gestalt Theory, describes why we inherently see some things, such as consistently aligned text, as easy to read, and inherently "right" and other things, such as inconsistent alignment, as inherently "off" or less pleasant to view. Gestalt explains why our gaze is drawn to certain content more than others, and how emphasis (and de-emphasis) may be achieved in data visualizations.

 

If you've taken a workshop with me, we likely discussed Gestalt via a subset of slides drawn from the slidedeck at left, developed by Gavin McMahon of Fassforward consulting. 

 

Want to learn more?

Visual Perception and Reading Accuracy

Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods. William S. Cleveland & Robert McGill Journal of the American Statistical Assoc, Vol. 79, No. 387. (Sep., 1984), pp. 531-554.

 

Cleveland and McGill's foundational 1984 article, cited above, on graphical perception applied research methods to determining what types of elements commonly found in graphs are the easiest for perceiving numeric differences.

 

In a workshop, we usually discuss this research using:

  • A visual/spacial perception activity (handout here)

  • A research summary page with examples (handout here)

  • Plus a look at various examples 

 

Want to learn more?

The Chart Junk Debate

Useful Junk? The Effects of Visual Embellishment on the Comprehension and Memorability of Charts. Scott Bateman, et al. In the Proceedings of ACM Conference on Human Factors in Computing Systems (CHI 2010). Atlanta, GA, USA. April 10-15, 2010. 2573-2582.

 

Bateman et al took a look at whether chart junk - graphical embellishments - decreased the understanding and memorability of graphs, thus diving head first into the chart junk debate. Stephen Few reviewed their research and found it lacking, but at least potentially prompting further exploration. This article is a good start to considering whether the chart junk debate is as much one of purpose as one of substance, and consideration of the role of data visualizations in increasing engagement and interest in the content they portray. Article also may be reflected on in terms of Gestalt theory.

 

Want to learn more?

  • Read Bateman et al's article here

  • Read Stephen Few's response here

Data Visualization Memorability

What Makes a Visualization Memorable?. Michelle A. Borkin, et al. IEEE Transactions on Visualization and Computer Graphics (Proceedings of InfoVis 2013), 2013.

 

Borkin et al's article, cited above, details a research study on what makes a visualization memorable. It identifies what makes ANY image memorable and does not take into consideration whether the viewer is remembering anyting about the data viz. I reference it because it is being widely used to justify questionable work. Be sure to read Few's rebuttal if you read the article.

 

Want to learn more?

  • Read Borkin et al's article here

  • Read Stephen Few's response here

Examples for Discussion

Video: What do jelly beans have to do with Museum visitors?

 

This video, from the Denver Museum of Natural Science, provides an example of an intriguing visualization creating a stacked bar chart with jelly beans.

 

  • In what ways does it align with the principles associated with Gestalt Theory?

  • Where does it fall on Cleveland and McGill's continuum?

  • Are exact values clear? Is the overall message clear?

  • Is it memorable? Why or why not? 

  • Is it successful as a data visualization? Why? Why not?

 

Want to learn more?

  • Read about these and two other videos from DMNS here

Interactive Visualization: Making Mountains Out of Molehills

 

We usually discuss David McCandless as hanging out at one end of the dataviz thinkers' spectrum due to his willingness to forsake accuracy for aesthetics at times. How did he do with this 2007 data visualization that looks at global media coverage of deadly topics (swine flu, killer bees) versus actual deaths? You'll need to click through to his site to try it out. Be sure to toggle the "Level patterns" and "Scale by..." options in the bottom right.

 

  • In what ways does it align with Gestalt Theory?

  • How does he handle the desire for a common scale?

  • Is it successful as a data visualization? Why? Why not?

 

Want to learn more?

  • Explore the original 2007 visualization here

  • Explore the 2014 update here*
    * What did he change? Did the direct labeling work?​

 

bottom of page