In higher education institutions, data-driven decision-making is becoming increasingly important, especially in the analysis of student performance. The aim of my research is to develop a grading analytics framework that aids in revealing the connections between subjects taught within a given program, as well as to automate this framework using a web application. These two interconnected topics target the comprehensive evaluation and improvement of the curriculum of higher education programs.

First, I describe how we built a grading analytics framework utilizing the results of students from the Business Informatics BSc program at Corvinus University of Budapest. In addition to the grades obtained in courses, we also considered the number of exam registrations and course enrollments, the gender of the students, and their year of study. As a result of the analysis, we developed a new performance metric that takes into account not only the final grade but also the pathway leading to it when evaluating student performance. Using Least Absolute Deviation regression modeling, we were able to determine how the results of individual subjects are influenced by courses taken in previous semesters. Based on the results, a curricular network can be established that builds on the actual relationships between courses and can help optimize subject prerequisite systems.

However, this framework required a lot of manual work and technical knowledge, so I worked on developing a platform that minimizes human intervention and optimizes the process. I developed the Grading Analytics application using the R programming language and the Shiny package, which provides an easy-to-use web interface for teachers and program coordinators. Previously, the analysis process took several months, primarily due to data cleaning tasks. The application now automatically performs these steps, eliminating the possibility of human error and significantly increasing the efficiency of the process. Additionally, I incorporated various error handling mechanisms during the statistical analyses, which also support the users’ work. By using the application, teachers and program coordinators can quickly and efficiently analyze the relationships between subjects, contributing to the improvement of the curriculum.

The grading analysis framework and its automation not only increase the efficiency of analysis processes but also enable a comprehensive evaluation and optimization of higher education programs based on student outcomes.

Keywords: Learning Analytics, Student Performance Analysis, Higher Education Curriculum, Web Application Development.