Visualization of Large Web Access Data Sets

Hao, Ming C. ; Garg, Pankaj ; Dayal, Umeshwar ; Machiraju, Vijay ; Cotting, Daniel

HP Development Company

关键词: visual clustering;    data access;    mass-spring;    relationships;    web transactions;   

RP-ID  :  HPL-2002-71

学科分类：计算机科学（综合）

美国|英语

来源: HP Labs

【 摘 要 】

Many real-world e-service applications require analyzing large volumes of transaction data to extract web access information. This paper describes Web Access Visualization (WAV) a system that visually associates the affinities and relationships of clients and URLs for large volumes of web transaction data. To date, many practical research projects have shown the usefulness of a physics-based mass-spring technique to layout data items with close relationships onto a graph. The WAV system: (1) maps transaction data items (clients, URLs) and their relationships to vertices, edges, and positions on a 3D spherical surface; (2) encapsulates a physics-based engine in a visual data analysis platform; and (3) employs various content sensitive visual techniques - linked multiple views, layered drill-down, and fade in/out - for interactive data analysis. We have applied this system to a web application to analyze web access patterns and trends. The web service quality has been greatly benefited from using the information provided by WAV. Notes: 5 Pages