Social Network Analysis is an emerging research topic and can be applied to a variety of domains. Modeling and visualizing any domain in the form of a network can be cumbersome if there are too many nodes. This is a major disadvantage as pattern recognition by humans becomes difficult. One feature which I particularly liked about the tool SocialAction is that users can change the measures on a network without causing the nodes to lose their orientation.The nodes can also be aggregated to reduce complexity or find subgroups. This interaction is very useful in exploratory data analysis.
It is interesting to know how the nodes are ordered and the criterion on which the nodes are ranked can be easily related to common tasks in interpreting social networks. The concept of ranking the nodes appears to be a good approach and the author's choice of using scatterplots to represent nodes that meet criteria across two rankings is justified. The matrix summary shown was hardly readable and failed to meet the intended purpose.
Not much is known about the algorithms Prefuse and Piccole used to generate the visualizations. The performance of the tool could have been gauged better if the complexities of the two algorthms were known or if SocialAction was compared with similar tools(if it exists). Overall, the idea of social network analysis tool based on the concept of node ranking was interesting.
0 comments:
Post a Comment