Multiscale visualization of relational databases using layered zoom trees and partial data cubes

Abstract

The analysis and exploration necessary to gain deep understanding of large databases demand an intuitive and informative human-computer interface. In this paper, we present a visualization system with a client-server architecture for multiscale visualization of relational databases. The visual interface on the client supports web-based remote access. We use zoom trees to represent the entire history of a zooming process that reveals multiscale details. Every path in a zoom tree represents a zoom path and every node in the tree can have an arbitrary number of subtrees to support arbitrary branching and backtracking. Zoom trees are seamlessly integrated with a table-based overview using "hyperlinks" embedded in the table. To support fast query processing on the server, we further develop efficient GPU-based parallel algorithms for online data cubing and CPU-based data clustering. Also, a user study was conducted to evaluate the effectiveness of our design.