Go beyond the traditional clustering and predictive models to identify patterns in your business data. Social network analysis describes customers' behavior, but not in terms of their individual attributes. Rather than basing models on static individual profiles, social network analysis depicts behavior in terms of how individuals relate to each other. In practical terms this approach highlights connections between individuals and organizations and how important they might be in viral effect throughout communities and particular groups. For business purposes, social network analysis can be employed to avoid churn, diffuse products and services, and detect fraud and abuse, among many other applications. This course shows you how to build networks from raw data and presents different approaches for analyzing your customers, focusing on their relationships and connections within the network.
Based on the recognition of customers', or organizations', roles within communities or special groups, you can improve business performance and better understand how your customers are using products and services. In addition to the network analysis approach to linking distinct entities, playing different roles on particular connections, this course also shows you a set of network optimization algorithms that you can use to solve a variety of complex business problems. Methods such as minimum-cost network flow, shortest path, linear assignment, minimum spanning tree, eigenvector, and transitive closure are presented in a business perspective for problem solving.
This course contains practical examples based on SAS Social Network Analysis Server and PROC OPTGRAPH.
- identify the type of the data and the nodes and roles in a network perspective
- identify the type of the data and the possible links between the actors within the network
- define the possible weight for nodes and links in a network perspective and the methods to build a network upon the data available, considering the distinguished importance of nodes and links within the network
- recognize the different types of groups and clusters of nodes based on their relationships within the network, such as communities, connected components, bi-connected components, core, cycle, and reach (ego) networks
- compute the network metrics such as degree, influence, closeness, betweenness, hub, authority, eigenvector, and clustering coefficient, and analyze the meaning of them, considering the business scenario, the industry involved, and the problem to solve
- perform network optimization based on several distinct algorithms like shortest path, minimum-cost network flow, linear assignment, eigenvector, minimum spanning tree, and transitive closure
- apply network analysis to solve real business problems in different industries.
Who Can Benefit
- Business analysts, statisticians, mathematicians, network engineers, computer scientists, data analysts, data scientists, quantitative analysts, data miners, marketing analysts, risk and fraud analysts, analytical model developers, and marketing modelers in all industries, including but not limited to communications and entertainment, banking and finance, insurance and retailers
- Before attending this course, you should have at least a beginner background in statistics and mathematics. You should also be minimally familiar with SAS programming.