Proposta de uma plataforma para análise de sentimentos nas interacções do Moodle durante o processo de ensino-aprendizado: Faculdade de Engenharia - UEM

Abstract

In recent decades, there has been an increase in the amount of information processed, stored, and made available in documents, mainly in the digital environment. This phenomenon has had a direct impact on the educational field. Among the various tools in virtual learning environments, the Discussion Forum stands out, as it enables the virtual mediation of relationships between students and teachers. Aiming to assist the Faculty of Engineering at UEM in monitoring student’s sentiments and supporting decision-making, this thesis proposes the creation of a platform that employs data science techniques, especially sentiment analysis in discussion forums. The study is based on the following question: to what extent can a Platform for Sentiment Analysis in Moodle Interactions during the teaching and learning process enable and streamline the educational process? For the development of the project, qualitative research was conducted, supported by a literature review and a survey directed at students from the Faculty of Engineering at UEM. Data science and programming techniques were combined to enable the analysis of students' sentiments. As a result, a dashboard was developed to present data visually through graphs and tables that summarize students' learning status. The analysis is based on sentiments extracted from natural language texts in Moodle discussion forums, which will then be categorized and labeled as positive, negative, or neutral using SentiLex-PT02. After the conclusion of the study, it was possible to determine that education can be enhanced through data science techniques, allowing for the quick and effective identification of students' sentiments in their comments and discussions. This enables teachers and administrators to gain valuable insights into pedagogical adjustments and the improvement of the teaching-learning process.