Qualiz: Gamification to Teach and Engage on the Integration of Requirements Engineering and Data Quality in Machine Learning

Abstract

The expansion of solutions based on artificial intelligence has intensified the reliance on machine learning (ML) techniques, making data quality a determining factor for the success of these systems. The effectiveness of models depends directly on complete, consistent, and representative datasets capable of supporting both the training and validation stages. However, the literature shows that issues related to poor data quality remain among the main causes of failures, biases, and performance degradation in ML applications. In this context, requirements engineering emerges as a strategic field, as it provides methods, processes, and guidelines that enable the identification of needs, the establishment of quality criteria, and the anticipation of risks associated with the collection, preparation, and maintenance of the data used in these models.
Despite the relevance of this topic, there is a significant gap in educational initiatives aimed at training professionals on data quality requirements for ML, especially in approaches that reconcile technical aspects with motivational strategies. Recent literature highlights gamification as a promising resource in this context, given its potential to make the learning process more dynamic, engaging, and effective. Within this context, this article presents Qualiz, a gamified prototype developed with the aim of disseminating good practices related to requirements engineering applied to data quality in ML systems. The prototype was made available to a group of 25 participants, who freely interacted with the platform and subsequently completed an evaluation questionnaire. The results revealed a predominantly positive evaluation: participants highlighted their good understanding of the topics presented, the relevance of the concepts addressed, and the adequacy of the quiz structure in facilitating the gradual assimilation of content. In addition, the interface was considered intuitive, and the playful elements were perceived as factors that increased interest and motivation during use. Based on these findings, Qualiz demonstrates significant potential as a complementary tool to support training in requirements engineering focused on data quality in ML, contributing both to the advancement of the field and to the education of professionals better prepared to operate in increasingly data-driven environments.