A Note about OPTGRAPH
SAS has delivered a very useful program in relation to social network analysis, called OPTGRAPH. This program includes a good range of calculations for network measurement, making it possible to compute metrics for distinct types of connections, such as the one used in the case study in Part II that is based on directed graphs, which are suitable for the telecommunications industry, and undirected graphs, which are suitable for most social network analysis scenarios, such as friendship, authorship, and employment, among others.
Considering both types of graphs, directed and undirected, OPTGRAPH provides algorithms to calculate degree—including in-degree (input connections) and out-degree (output connections)—eigenvector, closeness, betweenness, influence (first and second order), authority, hub, pagerank, and cluster coefficient. All these network measures can be calculated considering a particular weight or not. The weights can be any business attribute recognized as relevant for the social network analysis. In terms of the telecommunications environment, weights in relation to the nodes can be a customer’s value based on any corporate approach, a customer’s segmentation, the average customer’s billing, and so on. Weights in relation to the links can be the frequency of the calls, the average or total duration of the calls, the rate of the calls, the type of the calls (national, international, or local), and so on. Weights are useful to distinguish ...