Wednesday, November 16, 2011

Reaction-Balancing Systematic and Flexible Exploration of Social Networks

This paper was bit different than the other ones we have read so far. I liked the structure and organization of the paper. It starts with informing us about the gravity of the problem that why network analysis is a difficult task, how it is different from other analysis, what approach they followed to solve the problem.

The authors said that when it comes to network analysis the approach needs to be as flexible as systematic. I agree with the point that the analysis of a network by two different analysts presented with same data may lead to different conclusions unlike in biological analysis wherein both the conclusions should be same. Thus the exploration of social networks should be flexible. The user should have the power to select some interesting criteria and all the nodes should be arranged accordingly.

I liked the way they have discussed all the individual elements/principles used in coming up with their final solution, SocialAction. There has been in-depth discussion supported with diagrams and examples about Ranking & Coordinated views, Ranking social networks, filtering etc.

The authors said that when it comes to network analysis different measures can have different meaning in different networks. Thus they need something that allows user to examine the breadth of structural measures in a network. This is implemented by Ranking and coordinated Views. But then they that there could still be problems if the network is large as ordered list may be quite long and thus network layout may become illegible. They solve this problem in SocialAction by aggregating the nodes based on links, hence compressing the network and then examining the communities of interest rather than the individual elements.

One very good point underlined by this paper is that user perceptual clarity improves when the number of visualized elements is limited. So they have used the principle of Filtering after ranking the data which allows user to focus only on areas of his interest.

I can use the principles of ranking and filtering and some other approaches discussed in the paper in my class project and thus for me this paper has been really important and useful to read.