Behavior Represents Achievement: Academic Performance Analytics of Engineering Students via Campus Data

Behavior analysis is an important research direction in the field of social computing. Analysis, understanding and intervention of students' daily behavior patterns may provide a new to improve academic performance in addition to the classroom. This study, thus, uses campus behavior data to study the performance of undergraduate students. The experimental results manifested that behavior preferences were closely related to GPA. Specifically, GPA was positively correlated with library visit number (r = 0.386, p <; 0.001), the total number of books borrowed (r = 0.355, p <; 0.001), and the number of meals had in the school canteen. Morning activities in dorm was positively correlated with GPA while afternoon and night activities were negatively correlated with GPA. A Hierarchical linear regression model revealed that students' behavior in campus can explain GPA of adjusted R
²
of 0. 23, with meal habits 0.100 (p <; 0.001), daily schedules 0.116 (p <; 0.001), library visits and reading preferences 0.015 (p <; 0.001).