Identifying Relevant Users and Groups in the Context of Credit Analysis Based on Data from Twitter

In recent years several online social networks have emerged with very different purposes. This huge popularity is associated with common functionality to provide users with new ways to interact, producing content and commenting on various subjects and interests. This fact makes social networks favorable to research related to the organization and management of large amounts of data, besides constituting an ideal environment for knowledge extraction and application of data mining techniques. In this sense, this study collected data from the social network Twitter, and compared them with data from a financial institution in order to model the network and analyze their similarities. Three thousand users from Twitter were analyzed and 504 matched with the database from company for credit analysis. The results demonstrated that most of those users have more credit restriction than their neighbors, and users with no restrictions normally have also neighborhoods with no credit restriction as well.