This research is motivated by the suboptimal Technological Pedagogical Content Knowledge (TPACK) competency of Automotive Professional Teacher Education (PPG) students in designing meaningful, contextual, and technology-integrated learning. The use of Learning Management Systems (LMS) in the Student Understanding and Learning course is still limited to material distribution and assignment collection, thus not supporting data-based reflective, collaborative, and adaptive learning. This condition encourages the need to develop an andragogy-based digital learning system that utilizes Behavioral Learning Analytics to improve students' TPACK competency.This study aims to develop a valid, practical, and effective Smart Andragogical LMS–Padlet based on Behavioral Learning Analytics in improving the TPACK competency of Automotive PPG students. The research method used Design and Development Research (DDR) which includes the stages of model development, model validation, and model implementation. The development stage includes needs analysis, andragogy-based LMS architecture design, Padlet integration as a collaborative reflection medium, and the development of a learning analytics dashboard. Data collection was carried out through document review, interviews, questionnaires, TPACK pretest–posttest, and analysis of LMS and Padlet activity logs. Data were analyzed using thematic analysis, descriptive statistics, N-Gain test, and learning analytics analysis. The results showed that the developed product met the criteria of being valid, practical, and effective. The product consists of an andragogy-based LMS, a Padlet collaborative reflection space, a learning analytics dashboard, TPACK-based learning tools, and an evaluation instrument. Product implementation increased the average TPACK competency of students from 61.83 to 86.47 with an N-Gain value of 0.65 (moderate category). The student engagement level reached 93.86% (very high category), while the analytics dashboard was able to monitor the development of student competencies and engagement in real time, thus supporting data-based learning decision-making.
Copyrights © 2026