<![CDATA[Attendance monitoring and academic evaluation are two critical performance indicators in non-formal education institutions such as tutoring centres. HiveEdu, a tutoring centre in Manado, previously managed both processes using manual spreadsheet entries, which resulted in high administrative workload, frequent data entry errors, and delayed parental reporting. No analytical tool existed to detect early signs of declining student performance. This research develops a web-based e-raport system with a learning analytics module powered by Multiple Linear Regression (MLR). The system adopts a decoupled architecture using Next.js for the front-end, NestJS for the back-end, and PostgreSQL as the relational database. Security is enforced through JWT authentication and Role-Based Access Control (RBAC) with three roles: ADMIN, TEACHER, and USER. The MLR engine predicts each student's next exam score (Y) based on attendance percentage (X1), average tryout score (X2), and teacher objective score (X3). Predictions are categorised into risk levels (HIGH, MEDIUM, SAFE) and visualised in an Early Warning dashboard. Testing covered 60 black-box functional scenarios and 80 RBAC authorisation scenarios, all 140 yielding expected results. Technical validation using Mean Squared Error (MSE) on a synthetic dataset produced MSE = 73.32. All features including academic records, attendance tracking, analytics, early warning, PDF e-raport, and Excel export were successfully implemented and tested.]]>
Copyrights © 2026
