Basser Seminar Series

SGViewer: a faceted search interface that blends aggregate views with point distributional views that reveal clusters and outliers.

Speaker: Dr Mark Sifer
University of Wollongong

Time: Friday 6 August 2010, 4:00-5:00pm
Refreshments will be available from 3:30pm

Location: The University of Sydney, School of IT Building, Lecture Theatre (Room 123), Level 1

Add seminar to my diary

Abstract

An alternative to content based search such as that offered by Google, is faceted search which provides users with multiple directories (facets) to support top-down navigation through a corpus. While this approach is not workable on the open web because of the absence of consistent meta-data, it is applicable to large curated corpora.

This talk will present a faceted search system that has been integrated with a content-based search engine to support high recall queries that can return many relevant results. The system provides aggregated landscapes for facets onto which individual search results are rendered, allowing clusters and result outlier features to be seen. This is supported for both ordinal and categorical facets. The talk will cover the visualisation design, system architecture and a demonstration. The system can also be used for traditional OLAP data such as sales data, where search results are replaced by some ranking of individual sales. This will also be demonstrated.

The talk is based on a demonstration given recently at this years ACM SIGMOD in Indiana.

Speaker's biography

Mark Sifer is a lecturer with the University of Wollongong where he has been teaching software development and data analysis subjects since 2005. He has also taught at Sydney University and UTS and worked at a number of companies including IBM Research (Haifa, Israel) and CSC (Australia). He originally completed a BSc at UNSW and a PhD at UTS. His long standing interest has been in the design of user interfaces for the exploration of various kinds of complex data.