Generación automática de reportes de desempeño estudiantil basado en LLMs

Abstract

Abstract: This Curriculum Integration Project, entitled "Automatic Generation of Student Performance Reports Based on LLMs (DataEduca)," aims to develop a prototype that integrates data analytics and utilizes a large-scale LLM modeling language, specifically Gemma3:12b. The result is the automated generation of narrative reports in PDF format, detailing the performance, attendance, and academic achievement of each student. This system arose from the challenges faced by many educational institutions, which manage student data manually or, in some cases, with tools that are difficult to interpret, making effective student tracking challenging. The project adopts the Design Thinking methodology, effectively following each of its phases for the successful development of the prototype. Furthermore, this system aims to facilitate the understanding of each student's academic performance, enabling the identification of performance issues and providing recommendations on how to address them. This reduces the administrative burden and fosters effective, evidence-based pedagogical intervention. This system represents a significant contribution to the use of data analytics with artificial intelligence in educational contexts, with the potential to scale and adapt to other educational institutions.