Visual Analysis of Relation Patterns in Graphs and Text Corpora

PhD Thesis Proposal Defence


Title: "Visual Analysis of Relation Patterns in Graphs and Text Corpora"

by

Mr. Weiwei Cui


Abstract:

We are in the midst of a data explosion and the volume of information 
generated and stored continues to grow at an unprecedented rate. For 
example, Facebook has more than 300 million active users and Twitter 
publishes more than 300 new messages every second, and the numbers keep 
increasing. Exploring and analyzing this enormous amount of data become 
increasingly difficult. Information visualization can help analyze huge 
and complex data by turning them into visual representations to exploit 
the tremendous pattern-recognition capability of the human visual system. 
One important research problem in information visualization is to present 
relations, which can be found in many datasets from the real word such as 
social networks and news corpora. The existing relation visualization 
techniques, such as graphs and projections, are only good at showing 
simple patterns in relatively small datasets. In this thesis, we propose 
some advanced relation visualization techniques for large datasets (e.g., 
graphs with thousands of nodes) and complex relations (e.g., 
multi-relations and time-varying relations in news corpora).

This thesis is composed of three main parts, each of which addresses an 
important problem in relation visualization. In the first part, we deal 
with the visual clutter problem which significantly reduces the 
effectiveness of large graphs. We propose a geometry-based edge clustering 
technique that can group edges into bundles to reduce the overall edge 
crossings and reveal underlying patterns in large graphs. In the second 
part, we present an enhanced word cloud layout that keeps the semantic 
relations between the displayed words in a sequence of word clouds 
generated over time for dynamic document data. In the last part, TextWheel 
is introduced to visualize complex micro-macro relations within news 
streams. The effectiveness of these methods has been demonstrated through 
extensive experiments using both synthetic data and data from real 
applications.


Date:  			Tuesday, 11 January 2011

Time:           	2:00pm - 4:00pm

Venue:          	Room 3494
 			lifts 25/26

Committee Members:      Dr. Huamin Qu (Supervisor)
 			Dr. Chiew-Lan Tai (Chairperson)
 			Prof. Long Quan
 			Dr. Pedro Sander


**** ALL are Welcome ****