2025 ASEE Annual Conference & Exposition

The integration of Artificial Intelligence (AI) into educational tools has revolutionized modern education by enhancing pedagogical practices and learning analytics. The emergence of Large Language Models (LLMs) has further accelerated this transformation by enabling complex analysis of textual data that would otherwise be labor-intensive for instructors. Reflective writing is a key component in educational practices which foster deeper cognitive and metacognitive skills among students. Typically, reflective techniques require students to articulate their learning processes in natural language. However, the effectiveness of these practices is maximized when students receive feedback on their reflective writings. Due to the time-consuming nature of analyzing these writings, the implementation of reflective practices has been limited.
In this study, we introduce ‘Student-Reflect,’ an LLM-powered tool designed for the automated analysis of student reflections. Student-Reflect extracts students’ learning outcomes and challenges from their reflective submissions and visualizes the frequency distribution of these topics through a dynamic dashboard. This visualization enables instructors to apply timely interventions after each class session based on students’ learning trajectories. The analysis of the model's performance is promising, demonstrating over 95% accuracy in extracting meaningful topics for analyzing students' understanding of the subject matter.