A social network is a relational connection between individuals, including any relationships related to the exchange of information such as those among friends or colleagues. The most significant problem associated with social networks is establishing the role of key individuals, who are charged with conveying important messages among everyone involved. However, the most appropriate representatives of classes are difficult to identify. The aim of this study was to develop a method with which to facilitate the automatic pre-construction of a social network prior to any interaction and pre-identify representatives within the network. The goals of this research were: (1) construct a social network based on a self-organization maps and social network analysis, (2) verify differences between pre-constructed and actual social networks, and (3) identify key representatives and validate the efficiency of the social network.