CHAPTER 12
Visualization Capabilities for Social Network Analysis
Probably the most important stage of the social network analysis process is the data analysis, particular the results analysis, which comes from computing the network measures. One of the final stages of analysis is data visualization of the network structure and, therefore, getting some insights from the way the network behaves and mostly from the relations among the nodes.
There are several different approaches to analyzing the outcomes from a social network, which, in practice, becomes analyzing the outcome data. According to the business issue, such as churn, sales, marketing, risk, fraud, and others, different sets of steps can take place. In order to analyze the social network measures for churn, it is important to consider the most influential customers and how they relate to the other customers. Are they influencers? If they decide to leave, what is their likelihood to lead their pairs on this same path? Some particular centrality measures can direct the analysis toward to this main goal. First- and second-order centralities as well as eigenvector and some measures like influence based on page rank, authority, or hub can be deployed to define the best set of measures to identify central customers to be retained. However, if the subject is marketing and sales, a different set of network measures can be used to understand who the best customers are when those particular business issues are considered. Network ...
