Visualisation and Analysis of Large and Complex Networks

The Visualisation and Analysis of Large and Complex Networks (VALACON) project aims to design, implement and evaluate effective and efficiant visualization, analysis, interaction methods for large and complex networks, such as terrorist telephone call networks and biological networks.

The research is looking into how to define good visualisation for large and complex networks, how to model the problem of creating good visualisation as an optimisation problem, and how to solve the optimisation problems efficiently, effectively and elegantly.

The project addresses key scientific challenges including scalability, visual complexity and domain complexity.

Scalability refers to data complexity to deal with huge networks that have millions of nodes and edges;
Visual complexity refers to the complexity of understanding visualization of large and complex networks due to limited human percetion and congnition;
Domain complexity refers to the specialised knowledge required to define good visualization of domain specific networks, for example, a background in biology (sociology) is needed to understand a biological network (social network).

What will this research achieve?

A good visualisation of a large and complex network is worth more than millions of words. Visual analytic tools are needed in many areas. For instance sensor network companies use visualisation to find the position of a sensor by visualising the distance to other sensors with known positions. Biotechnology companies use visualisation to show many complex chains of interaction such as metabolic pathways. And law enforcement authorities use social networks to identify groups involved in, for instance, insurance fraud, through the links between them such as phone calls.

Who will benefit?

The beneficiaries include intelligence analysts, biologists, life science researchers, sociologists, law enforcement officers and epidemiologists. The project is also generating ideas, methods and systems for industry and a pool of PhD graduates and researchers.

What are the key features?

The team has skills in visualisation and analysis, based on knowledge of algorithms, optimisation, psychology and graphics
With much longer experience in this area than any of the new groups, we currently lead the research community
We work with domain experts and collaborate with world experts
We convert a visualisation problem to a mathematical problem, which is usually an optimisation problem such as finding the maximal value with given constraints. The solution of this mathematical problem can then be converted back to a solution for the visualisation problem.
We develop techniques that create visualisations of networks that contain millions of nodes
We provide navigation methods for effective and efficient navigation of large data sets
We integrate visualisation with analytical methods and tools for data mining
This extends to support three-dimensional visualisation.

Progress update

Research team

Publications

Visualisation and Analysis of Large and Complex Networks

Progress (Research Outcomes)

1. Research Excellence

We have been dominating winners’ platforms in international competitions, winning IEEE InfoVis 2004, GD 2005 and GD 2006
Project leader Seokhee Hong won the CORE Chris Wallace award 2006 for her research contribution on Theory and Practice of Graph Drawing. This is the annual award for the best young Australian researcher in Computer Science.
We designed, implemented and evaluated the visual analysis of large and complex clustered graphs
We designed, implemented and evaluated the visual analysis of large and complex hierarchical graphs
We designed, implemented and evaluated the visual analysis of large and complex scale-free networks
We designed, implemented and evaluated the visual analysis of large and complex small world networks
We designed, implemented and evaluated the visual analysis of large and complex temporal networks
We designed, implemented and evaluated the visual analysis of large and complex dynamic networks
We designed, implemented and evaluated the visual analysis of large and complex sensor networks
We designed, implemented and evaluated the visual analysis of large and complex biological networks
We designed, implemented and evaluated the visual analysis of large and complex social networks
We designed, implemented and evaluated the visual analysis of large and complex transportation networks
Developed a prototype visual analysis tool GEOMI

2. Collaboration

3. Education

Designing and delivering courses at the University of Sydney in Information Visualisation and Computational Geometry
PhD education: 10 students

4. Commercialisation

5. Service to Community

Running an ARC Research network EII Task Force Team on Network Analysis and Visualisation
Organising NICTA workshops on social networks and biological networks
Organising international conferences: APVIS 2005, APVIS 2007

Research team

Project leader:

Seokhee Hong

Research staff:

Kai Xu,

Xiaoyan Fu

PhD students:

Adel Ahmed,

Damian Merrick,

Colin Murray,

David Fung,

Weidong (Tony) Huang,