User Tools

Site Tools


Visual Analytics Course


Visualization is about perceiving features and structure in images. It is also about giving the user control over the data to analyze. As a consequence, visualization naturally links with perception issues – as well as with cognition.

Visualization interfaces fall within a category Ben Shneiderman called “Direct Manipulation Interfaces”, to refer to systems having the following properties:

  1. Continuous representation of the object of interest.
  2. Physical actions or labeled button presses instead of complex syntax.
  3. Rapid incremental reversible operations whose impact on the object of interest is immediately visible. (Shneiderman, 1982, p. 251)

Citing Hutchins et al., we can only agree with the fact that […] the feeling of directness is inversely proportional to the amount of cognitive effort it takes to manipulate and evaluate a system […]. Designing and building a visualization aims at helping users understand and interpret data, from which they can access information and potentially find new knowledge. Obstacles are on the way, as in any communication channel.

Obstacles roughly reside at three different levels, each inducing distance between the author's intention (the visualization designer) and the reader's interpretation:

  • Operational level/distance. Sometimes referred to as the “readability” of a visualization (or more generally of a message). The ease of, or the amount of efforts a user must deploy in order to perceive all elements of a visualization. The operational level is concerned with the lexicons (elementary elements used to form the image or the message) and their characteristics that can affect/alter/help perception. These elements are there to improve readability, suggest where information is, but may unfortunately disturb the user from the author's intention.
  • Articulatory level/distance. Articulatory distance concerns the relationship between the meanings of the expression and their physical form. Physical form can be a sequence of keystrokes or mouse movements and clicks. This is where structure is, as far as the author is concerned. Lexicons organize into higher level elements to bring in structure and organize information. This is also here that navigation is suggested and operated by users. Visualization will typically rely on specific ways to organize visual elements on the screen, and offer interactino devices to users so they can “play” with the view in order to gain insight on the displayed information.
  • Semantic level/distance. Semantic distance is the relationship between what the author wants to communicate and the meaning of the corresponding expression in the visualization. When concepts are represented directly in the visualization, semantic distance is decreased. When complex, ambiguous or vague icons or words are used to represent a concept, semantic distance is increased. Semantic distance resides on top of the previous two. This is where meaning enters the scene. Colors may bear some specific meaning – this may depend on cultural conventions for instance. Shapes may induce interpretation, labels, etc. A combination of position, colors, and fonts may send a message that goes the opposite way to what authors intended.

(Italicized definitions are borrowed and adapted from Barbara Sheehan et al. (2009). AMIA Annual Symposium, pp. 584–588.)

Conversely, we can look at the distance the other way round and try to see whether a user can efficiently use a system. Hutcins et al. usefully consider different definitions for what they call the semantic and articulatory distances:

  • Semantic distance concerns the relation of the meaning of an expression in the interface language to what the user wants to say. Two important questions about semantic distance are
    1. Is it possible to say what one wants to say in this language? That is, does the language support the user’s conception of the task domain? Does it encode the concepts and distinctions in the domain in the same way that the user thinks about them?
    2. Can the thing of interest be said concisely? Can the user say what is wanted in a straightforward fashion, or must the user construct a complicated expression to do what appears in the user’s thoughts as

a conceptually simple piece of work?

  • Where semantic distance has to do with the relationship between user’s intentions and meanings of expressions, articulatory distance has to do with the relationship between the meanings of expressions and their physical form. On the input side, the form may be a sequence of character-selecting key presses for a command language interface, the movement of a mouse and the associated “mouse clicks” in a pointing device interface, or a phonetic string in a speech interface. On the output side, the form might be a string of characters, a change in an iconic representation, or variation in an auditory signal.

Have a look at these examples.

Papers worth looking at:

InfoVis: a few definitions

The use of computer-supported, interactive, visual representations of abstract data to amplify cognition. [Card, S. K., J. D. Mackinlay, B. Shneiderman (1999). Readings in Information Visualization. San Francisco, Morgan Kaufmann Publishers.]

Information visualization utilizes computer graphics and interaction to assist humans in solving problems. [Purchase, H., N. Andrienko, et al. (2008). Theoretical Foundations of Information Visualization Information Visualization, Lecture Notes in Computer Science 4950, Springer Berlin, pp. 46-64.]

Read more on the InfoVis Wiki.

Visual features that matter

Think of the eye as a computing machinery. According to Ware, 40% of our cortical activity is devoted to processing visual signals. But this doesn't go without making errors. Care must be taken when designing an image to avoid the eye to misinterpret the information we want the image to carry.

Visual illusions are convincing of our visual system's flaws in interpreting visual signals.

Pre-attentive signals

Some signals do not rely on higher cognitive processes, but are processed immediately. Some other signals rely on higher, and lengthier, processes.

Patrick Healey's website on perception is a useful resource for understanding and experimenting with these ideas.

  • Pre-attentive features
  • Change blindness

Color must be used with care. Cartographers have been using color schemes for decades to map quantities and/or categories. Simple color schemes include the rainbow spectrum, red-blue or red-green ramps, and the grey-red saturation scale. The use of colors must carefully seek for (borrowed from Healey):

  • perceptual balance: a unit step anywhere along the color scale produces a perceptually uniform difference in color,
  • distinguishability: within a discrete collection of colors, every color is equally distinguishable from all the others (i.e., no specific color is “easier” or “harder” to identify), and
  • flexibility: colors can be selected from any part of color space (e.g., the selection technique is not restricted to only greens, or only reds and blues).

Cynthia Brewer's ColorBrewer is a useful and authoritative source to help designers choose among all possible and most readable color schemes.

A few design principles help choose the right way to go with colors.

Exercises / Assignments

  1. Go to the ColorBrewer webpage and experiment.
  2. Browse the web, and/or books, and build list of recommendations for visualization designers. A kind of “must”s and “don't”s.
  3. Go to the Visual Complexity website, select a visualization and discuss its features, be critical, point at its flaws, congratulate its advantages and nice properties. Organize your discussion around the three levels/distances: operational, articulatory, semantic.
  4. Go to the Cartastrophe website and learn from bad cartographic design.
/net/html/perso/melancon/Visual_Analytics_Course/data/pages/perception.txt · Last modified: 2013/10/02 17:38 by melancon