We introduce the notion of an Active Proxy interface, i.e. tangible models as proxies for physical data referents, supporting interactive exploration of data through active manipulation. We realise an active proxy data visualisation system, "MarioChart", using robot carts relocating themselves on a tabletop, e.g., to align with their data referents in a map or other visual layout. We consider a casual-data exploration scenario involving a multivariate campus sustainability dataset, using scale models as proxies for their physical building data referents. Our empirical study (n=12) compares active proxy use with conventional tablet interaction, finding that our active proxy system enhances short-term spatial memory of data and enables faster completion of certain data analytic tasks. It shows no significant differences compared to traditional touch-screens in long-term memory, physical fatigue, mental workload, or user engagement. Our study offers an initial baseline for active proxy techniques and advances understanding of tangible interfaces in situated data visualisation.

翻译：我们提出了“主动代理界面”的概念，即作为物理数据指代物代理的实体模型，通过主动操控支持数据的交互式探索。我们实现了一个主动代理数据可视化系统“MarioChart”，该系统利用在桌面上自主移动的机器人小车（例如，使其与地图或其他视觉布局中的数据指代物对齐）。我们考虑了一个涉及多变量校园可持续性数据集的休闲数据探索场景，使用比例模型作为其物理建筑数据指代物的代理。我们的实证研究（n=12）比较了主动代理使用与传统平板交互，发现我们的主动代理系统增强了数据的短期空间记忆，并能够更快地完成某些数据分析任务。与传统触摸屏相比，在长期记忆、身体疲劳、心理负荷或用户参与度方面未显示出显著差异。我们的研究为主动代理技术提供了初步基准，并推进了对情境化数据可视化中实体界面的理解。