Approaches and cognitive resources employed by skilled interpreters of visualizations

(most recent update 24 january 2018) (return to workshop front page)

The workshop opened with an icebreaker activity designed to pull forth the group's awareness of what approaches and cognitive resources they themselves draw on when interpreting visualizations. Working in pairs, each participant tried to interpret a visualization they had never seen before, from a domain outside their expertise, while speaking aloud their thoughts to a partner who was familiar with the visualization.

Overarching insights emerging from this activity:

• In addition to the information that is present in the representation itself, interpreters call on two types of prior knowledge:
• factual knowledge about the domain which is the referent of the visualization, and
• procedural knowledge of how to make meaning from visualizations.
• Over the course of an educational trajectory or lifetime, these two types of knowledge co-develop. The trick in skillful interpreting is not just to have these two knowledges, but also to retrieve them at the appropriate time and coordinate between them.
• If we are already basically familiar with a visualization type (e.g. the weather forecast that we view habitually), we "see through" the visualization and pull forth information about the referent without conscious effort.
• But if we are not already familiar with the visualization type (as in this activity), we fall back on a set of domain-general strategies that are retrieved and implemented with conscious effort. Figuring out what these domain-general strategies are, and how to foster and assess them, is at the heart of the work of this workshop.
• Different strategies are needed for dynamic or interactive visualizations as opposed to static visualizations. This workshop mostly dealt with static visualizations; the same issues should be re-visited for non-static visualizations.
• The visualization creators in our workshop group found it hard to refrain from talking about how to improve the visualization, rather than about how to tackle the process of interpreting the visualization as it exists. We will probably also have that problem in our role as instructors.
• Visualization-creators make assumptions about what the interpreter will know (e.g. what arrows signify) or how the interpreter will behave (e.g. click every link)--and such assumptions may be unjustified.
• A visualization that works for one audience may not work for another audience.
• No matter how experienced or expert we are in our primary field, we all become more novice-like when we venture outside our domain--as in an interdisciplinary collaboration.

Approaches and resources that we used while interpreting:

Different visualizations called forth different strategies, but some flexible, re-usable strategies emerged:

• First get the "gist" of the visualization.
• Search for memory of similar images. Classify it, if possible, into a known template (e.g. graph, map, molecule).
• Attend to the legend, caption, axis labels.
• Get a handle on what is the topic or theme of the map. Retrieve whatever big ideas you may have about that topic or theme.
• Figure out the scale
• Deploy strategies specific to that visualization template (graph, map, etc.)
• For map, discern what area is covered
• For graph, identify what the axes are
• Leverage general conventions, but with caution:
• Arrows may signify motion (or not)
• Recognize that color is often used to convey meaning, but:
• the same color means different things in different domains (e.g red means oxygen in chemistry and warm in oceanography).
• color fails to do its job when the viewer is color-blind or the image is reproduced in B&W.
• color is sometimes there just for decoration and may in fact have no significance.
• Don't be wedded to your first interpretation; consider alternative interpretations (yes, I could be looking at another planet, but perhaps it could also be the bottom of ocean....)
• Consider the motives of the visualization creator: What story was he/she trying to tell? Are they trying to manipulate me? Might there be biases embedded in the visualization?

What was not articulated:

In thinking back over this activity, the workshop conveners note that none of participants talked about how exactly they went about the process of turning perceived plus retrieved information into new meaning. They talked about what leads up to that step (e.g. figure out the scale, retrieve existing ideas about the topic). And they talked about what should follow that step (e.g. consider alternatives to your first interpretation). But the actual creation of new understanding about the referent seems to have happened through unarticulated processes. This may be our biggest challenge in educating skillful visualizers.