Natural User Interface Design in DA-TU: An Interactive Clustered Data Visualization System

Clustered Graph Visualization, such as DA-TU [2], has been successfully applied in the field of large scale relational data visualization for data analytics. However, it was very inefficient in interactions by using WIMP based HCI in DA-TU, in which a Control Panel was required to swap among eight different modes of operation. In this paper, we attempt to use a new multi-touch enabled NUI method for supporting interactions in DA-TU. We aim to increase the efficiency of HCI process in DA-TU. To achieve this, we first transplanted DA-TU from WIMP based PC to NUI based iPad. We then designed a set of new multi-finger gesture vocabularies to support DA-TU navigations in tablet environment. Adopting such set of rich HCI vocabularies achieves better efficiency and user experience in data exploration. We have conducted usability studies that demonstrated our hypothesis.