Click or not: different mouseover effects may affect clicking-through rate while browsing interactive information visualization

Keyuan Zhou, Wenjie Wu, Junhan Zhao, Mingran Li, Zhenyu Cheryl Qian, Yingjie Victor Chen

Abstract: Given the thriving of information visualization systems, rich interactivity is expected and essential for users to acquire further and insightful information. Mouseover has been widely used to hint interactivity or details about the hovered elements. However, in a mouseover-click visualization system we designed based on a hierarchical Sankey diagram, we have observed that a mouseover effect interfered users from clicking items and hence reduced the efficacy of the interaction and hid the information in the next levels. This paper examined this situation and studied the effectiveness of different mouseover hints on affordance for inviting clicks. We identified four types of perceived affordances and created seven designs, adopting Mechanical Turk to measure the effectiveness. Results showed that five out of the seven designs significantly improved users’ click-through rate, whereas more than 60% of users failed to click the hovered elements for revealing the next level of information without proper affordance hints, even after they brushed over these elements for multiple times. By comparing the performance of different designs, we proposed guidelines that support the reliability and motivate the interactivity of visualization design.