Proposta de solução para apoiar o processo de análise de dados textuais de avaliações institucionais

Abstract

This study developed and evaluated a decision-support tool for academic managers, utilizing NLP for textual data analysis and information visualization techniques. In the academic context, interpreting qualitative data generated from responses to the CPA (Comissão Própria de Avaliação) self-assessment questionnaire poses significant challenges due to the complexity and volume of this data. To address these challenges, the DSRM (Design Science Research Methodology) methodology was applied across three development cycles, resulting in the artifact Analisa CPA, featuring functionalities tailored to the context of institutional management. The artifact identified the sentiments and emotions of comments, in addition to providing data visualizations adapted to different manager profiles, such as directors and members of the Rectory and the CPA. An evaluation through interviews with managers validated the positive impact of the solution, demonstrating that data-driven tools enhance decision-making in this environment. It is concluded that the tool facilitates the decision-making process, offering an accessible and functional interface that meets the needs of academic managers.