Nowadays, there are many methods proposed for helping users to efficiently find relevant contents on the Internet. However, there is still no guarantee for users to be always able to develop an appropriate perception/understanding toward the topics they concern in an efficient manner. To solve the above problems, we propose a novel visualization system which aims to improve users’ perception efficiently through both audio and visual cues. Through our system, users can explore and manipulate data in an interactive and multi-resolutional manner. In addition, by integrating the concept of harmonic chords, our system auditorily represents certain structures of multiple data clusters chosen by users. Furthermore, our system highlights groups of related data in the information graph dynamically. This approach can help user to serendipitously find hidden data cluster with similar pattern of relationship.