Cognitively inspired and perceptually salient graphic displays for efficient spatial inference making

Abstract

Developing a visual hierarchy in map displays that is congruent with thematic levels of relevance is a fundamental cartographic design task. Cartographers employ a set of visual variables (e.g., size, color hue, color value, orientation, etc.) for 2-D, static maps to systematically match levels of thematically relevant information to a perceptual hierarchy based on figure–ground relationships. In this article, we empirically investigate the relationship of thematic relevance and perceptual salience in static weather map displays. We are particularly interested in how novices’ viewing patterns are modified when thematically relevant items are made perceptually more salient through design. In essence, we are asking whether perceptually salient elements draw novice viewers’ attention to thematically relevant information, whether or not users have domain knowledge.
In a factorial experiment, we ask novice participants to evaluate the wind direction in weather maps before and after training all participants on meteorological principles. Our empirical results suggest that display design (i.e., saliency) does not influence the accuracy of response, whether participants have prior knowledge or not
(i.e., training). Analysis of the eye-movement patterns, however, suggests that display design does affect viewing
behavior and response time. These findings provide rare empirical evidence for generally accepted design practices
within the cartographic community (e.g., the effects of visual variables). We chose weather map displays as one
typical example of commonly used maps for our study, but the methods employed are generic enough to be applicable to any spatial display (static or interactive) that might be produced by GIScientists, cartographers, geographic information system (GIS) practitioners, and others. Key Words: empirical study, eye-movement analysis, geographic visualization, spatial inference, weather maps.