Social media has found its way into education and, together with team formation, has started to play a significant role in students' university progress. Previous research has tried to generally analyze and give predictions about the influence of social media on students' learning curve but was not concentrated on understanding students' behavior within teams and directly linking it to Belbin roles, which is crucial for forming teams. In this paper, we are working with real-life data extracted from official university channels will allow us to propose a methodology and a pilot Belbin role automation tool to look further into the specifics of the problem. In addition, linking this to team roles and behavior concerns within the project teams will open the horizon for further research on how the performance of students within the teams is affected. We propose to create a primary tool and framework for validating Belbin roles through real-life social network data. Results show that it is possible to identify Belbin's roles using natural language processing. The future work direction is to refine the words associated with every personality trait.