Motion in casual InfoVis and the interrelation between personality, performance, and preference

Current real-world casual information visualization systems use motion to presumably increase system appeal and repeated use. Despite empirical evidence suggesting that motion is distracting and not particularly good for data encoding, its continued use may indicate benefits independent of traditional utility measures. We present two experiments designed to highlight differences and disparities between task performance, perceived system utility, and subject preference for static and moving glyphs. A music library visualization provided a casual, practical, nuanced, and pertinent domain application. Song glyphs represented beats per minute and beat strength using a line encircling the glyph. Beat encodings used one of four methods: a static sinusoidal pattern (static), an oscillating movement pattern (motion), an oscillating sinusoidal line (redundant), and a sinusoidal line that moves to the currently playing song (extraneous). Subjects first performed a pair-wise glyph comparison task to identify conditional differences in speed and accuracy. Then, a more realistic playlist generation task was performed using these beat encodings to explore opinions and behavioral patterns using a qualitative approach. Our results show clear performance decreases associated with motion use in the comparison task. Despite this, more than half the subjects preferred a moving glyph to the static encoding. We identified three groups of subjects (utility, fun, and intuitive) where subjects were consistent within groups but distinctly different between groups. This suggests that casual visualizations could be customizable based on high level groupings, where pragmatism can be emphasized for some users and aesthetics and style can be enhanced for others.