学术文献库

As data visualizations have been increasingly applied in mass communication, designers often seek to grasp viewers immediately and motivate them to read more. Such goals, as suggested by previous research, are closely associated with the activation of emotion, namely affective arousal. Given this motivation, this work takes initial steps toward understanding the arousal-related factors in data visualization design. We collected a corpus of 265 data visualizations and conducted a crowdsourcing study with 184 participants during which the participants were asked to rate the affective arousal elicited by data visualization design (all texts were blurred to exclude the influence of semantics) and provide their reasons. Based on the collected data, first, we identified a set of arousal-related design features by analyzing user comments qualitatively. Then, we mapped these features to computable variables and constructed regression models to infer which features are significant contributors to affective arousal quantitatively. Through this exploratory study, we finally identified four design features (e.g., colorfulness, the number of different visual channels) cross-validated as important features correlated with affective arousal.