Table-Scale Model to Visualize Very Large Data

Introduction

Analyzing large amounts of data is difficult because of the limitation of screen space and human’s perceptual processing. Breakdown data analysis comes from information foraging theory, which guides users’ progress from summary information to detailed data by maximizing gains of valuable information per unit cost. In a tabular data set, aggregation is to group or combine tuples and attributes into summaries using various operators. In reverse, breakdown is to segragate summaries progressively into individual tuples. Iteratively applying aggregation and breakdown generates tables with diverse scales on different levels of abstraction, which form an aggregation space. The conceptual table-scale model offers a general framework. If visualized, the framework supports creation and navigation through the tabular space by providing an overview of all possibilities, visual representations of breakdown results and aggregation paths.

Documents

Qing Li, Chris North, A Survey on Aggregation Strategies in Information Visualization, submitted to Information Visualization Journal, 2005

Qing Li, “Table-Scale Framework for Navigating Large Tabular Data”, PhD Proposal, 2005