Tuesday, November 15, 2011

Reaction: Balancing Systematic and Flexible Exploration of Social Networks

This paper focuses on understanding and analyzing the complex social networks.  Today social networks hold possibly much more information than any of us can imagine. Exploring and mining relational data from such social networks is extremely challenging as the networks are themselves extremely huge with many nodes and interconnecting links and lack of appropriate methods to explore and analyze data from humongous networks.

This paper introduces and explains the tool called Social Action which is built with the aim of making the process of visualizing social networks and data exploration simple. Social Action gives the user power to give ranks to nodes in the network using two types of structures - ordered lists and a node diagram. The tool provides excellent functionality to the users to visualize the data with respect to ranking using colors. The nodes can be ranked into ordered lists by using different pre-defined criteria which rests upon user discretion. The different criteria helps the user interpret the rank of the node value with respect to other nodes in the group. 

SocialAction also allows the user to zoom in/out of the node graph and highlight and focus on only those nodes which are important. This makes the graph manageable and easy to interpret. I personally feel that the functionality of visualizing the group of nodes as sub-groups is a very good addition by the authors. This lets the user visualize complex graphs in a better way by grouping the nodes together which have a relation between them. This in turn allows the user to analyze the relation between such closely related groups and the ranking of such groups which I think is very important and has multiple applications. For example; In facebook, we always get friend suggestions - these people are related to one particular node (friend) in my social graph. But, the his/her subgroup might be closely related to mine and thus; the suggestion. This might be far-fetched and involve huge amount of processing but its a general idea.

Another useful functionality mentioned is highlighting of parts of graph which are based on particular value of user input which in normal condition involve re-arranging of the graph and thus disturb the user perspective vision. 
Overall, it is an excellent tool with great functionality which can be used on enterprise level and can be made better.