Applying Natural Language Processing to Teamwork - A New Dashboard for CTMTC Methodology

In our current society the acquisition of competences such as teamwork is essential. However, the evaluation of how this competence is developed is not easy and requires methodologies and tools to support the assessment process. In this sense several Learning Analytics tools have been developed. They explore students' interactions in different types of tools such as forums or instant messaging apps. However those tools are especially focused on the quantitative evaluation of the interaction and are not very usable. This work presents a new dashboard that analyzes students' Telegram interactions while they work as a team to address a project. The innovation of this tool lies in the functionalities included to explore not only numbers about messages, replies, type of messages, characters, etc., but the content of the texts. To do so natural language processing and sentiment analysis libraries were used. The tool has been tested successfully with 4 subject editions in which it is possible to appreciate an evolution in students' interactions.