Abstract
This paper introduces Tribefinder, a novel system able to reveal Twitter users’ tribal affiliations. Tribefinder establishes to which tribes individuals belong through the analysis of their tweets and the comparison of their vocabulary. These tribal vocabularies are previously generated based on the vocabulary of tribal influencers and leaders selected using Tribecreator. To demonstrate its functionality, in the case presented in this paper, the system was calibrated in three specific tribal macro-categories: alternative reality, lifestyle, and recreation. Apart from describing the methodology we used to create this system, we also provide some practical examples of its use, thus giving a first indication of its potential. Finally, we present the results of the adoption of a t-SNE visualization approach, useful to verify whether tribe members cluster closely together.