Log on:
Powered by Elgg

Kasturi Chaterjee :: Blog :: PIRE Proposal

May 13, 2008

Student Name: 

Kasturi Chatterjee

Supervisor’s Name and Title at FIU/FAU:

Dr. Shu-Ching Chen, Associate Professor

Name of the PIRE International Partner’s Institution:

IBM Research, China Research Lab

Supervisor’s Name and Title at the PIRE International Partner’s Institution:

Dr. Shixia Liu, Research Scientist, Information Visualization, IBM-CRL

Project Title:

Proposing a Graph Similarity Toolkit for improving Social Network Layout

Problem Statement:

Visualization is a cognitive process to develop images and mappings of objects in one’s mind. In computer and information sciences it is the visual representation of a domain space using graphics, images, animated sequences and sound augmentation to present the data, structure and dynamic behavior of large, complex data sets that represent systems, events, processes, objects and concepts. In this joint project, we will work on Information Visualization utilizing Automatic Graph Layout concepts. A detail implementation plan and algorithm for automatic graph layout with a special concentration on graph similarity will be recommended after a careful survey of the existing works on graph layout techniques.

Motivation and Impact:

Visualization have immense impact on various genres mainly to fields which deal with a huge amount of data. Apprppriate visualization techniques are capable of extracting many hidden interpretations of critical data which might be hard to detect under ordinary circumstances. Using dynamic graphs is one of the most popular techniques of data visualization specially for applications like Social Networks. Finding interesting subgroups of sub-graphs from a Social Network Graphical structure using graph similarity techniques is a very interesting topic and would help to design and analyze Social Networks structures better. 

Current Status:

This is a completely new venture both on my part as well as for the research group for IBM-CRL. They had been more concerned with the visualizaton aspects of the graph and want to extend it to the data analysis part as well using similarity metrics. Thus, we need to start building up the knowledge base, design and implement the proposed idea.

Research Roadmap:

In 2008, the work will be decomposed into the following major subtasks:

1.      Explore all the existing graph layout algorithms from multiple resources from academic, commercial products and open sources with a special concentration on Graph Similarity and Distance.

2.      Analyze the characteristics (advantages, disadvantages) of each layout algorithm.

3.      Define the research agenda of automatic graph layout recommendation for the team, and divide research responsibility with other colleague(s) of the Research team with the following objective:

                              a.       Recommend the layout algorithm for each input graph

b.      Recommend the layout algorithm for each sub-graph in the input graph



4.      Prepare a final document of the findings with experimental results and details of the recommended structure.


Relation to PIRE Core Research Projects:

Though information visualization is currently not directly related to the PIRE projects but there are ample future oppurtunities to extend our research in this field. For example, visualizing the post hurricane information as graph structures and analyzing the data and categorizing them using the similarity toolkit might be just one of the many implementations of my research at IBM-CRL. 

Keywords: kasturi, pire, proposal

Posted by Kasturi Chaterjee

You must be logged in to post a comment.