IEEE TCDL Bulletin
 
space space

TCDL Bulletin
Current 2005
Volume 2   Issue 1

 

Supporting Collection Development Decisions by Mining and Analyzing Digital Archive Usage Data

Jewel Ward†, Johan Bollen‡, Jeffrey Pearson†, Shing-Cheung Chan†, Hui-Hsien Chi†, Marie Chi†, Kristine Guevara†, Hsiao-han Huang†, Genesan Kim†, Maks Krivokon†, Bo H. Lee†, Pei-Han Li†, Fenny Muliawan†, Vu Nguyen†, Barry W. Boehm†, A. Winsor Brown†, Edward Colbert†, Alex Lam†, Mayur Patel†

†University of Southern California
Los Angeles, CA 90089
{jewelw, jpearson, shingchc, hchi, mariechi,
kguevara, hsiaohah, genesank, krivokon, bohlee,
peili, muliawan, nguyenvu, boehm, awbrown, ecolbert,
alexankl, mayurkup}@usc.edu
‡Old Dominion University
Norfolk , VA 23529
jbollen@cs.odu.edu
 

This 3D data mining tool is a prototype "decision support tool" to aid in collection development by combining the results from mining key event data with a 3D viewer to reveal the structure of relationships among retrieved items from registered patterns of user retrieval. The tool provides three functions: the generation of object similarities, the hierarchical clustering of a similarity graph and a Hyperbolic 3D visualization of the object tree.

The hierarchical clustering of object similarities reveals the high level structure of the object collection. The Hyperbolic 3D layout algorithm provides a visualization of the analysis results for extremely large volumes of data, while retaining view manageability and avoiding clutter. The interface is extended by a set of object lists that reflects the current location in the tree and aids navigation within and between tree levels. The search capability allows users to locate a particular object located within the tree structure. Our preliminary analysis indicates this tool can provide information about collections that standard usage reports do not. The information derived from the analysis is useful both for future collection development and for the creation of virtual collections.

Future steps include: performing a full analysis, determining what isn’t being looked at, incorporating a sort function, allowing the customization of log input attributes, creating a web-based version, exploring other visualization methods, visualizing the results by collection, and creating a web service that ties the output metadata/URLs to renditions. The software and documentation can be found at: http://greenbay.usc.edu/csci577/spring2005/projects/team7/.

Thumbnail of a poster from JCDL 2005

For a larger view, click here.

Thumbnail of a poster from JCDL 2005

For a larger view, click here.

Thumbnail image of poster

For a larger view, click here.

 

© Copyright 2005 Jewel Ward, Johan Bollen, Jeffrey Pearson, Shing-Cheung Chan, Hui-Hsien Chi, Marie Chi, Kristine Guevara, Hsiao-han Huang, Genesan Kim, Maks Krivokon, Bo H. Lee, Pei-Han Li, Fenny Muliawan, Vu Nguyen, Barry W. Boehm, A. Winsor Brown, Edward Colbert, Alex Lam, Mayur Patel
Some or all of these materials were previously published in the Proceedings of the 5th ACM/IEEE-CS Joint Conference on Digital libraries, ACM 1-58113-876-8/05/0006.

Top | Contents
Previous Article
Next Article
Home | E-mail the Editor