Abstract: The field of education has been revolutionized by advancements in big data techniques, enabling educators to gain precise and timely insights into students behavioral patterns. This newfound capability is invaluable for identifying specific student groups that require targeted attention and transitioning from relying solely on qualitative empirical knowledge to incorporating scientific quantitative analysis in student affairs management. To fully harness the potential of this revolution, a meticulously developed system was implemented to apply data mining clustering method in analyzing the campus network behavior of 3,245 students in a particular grade at B University. Over a careful four-year period, a comprehensive dataset comprising 23.843 million Internet access records was collected. Through thorough analysis, it was discovered that the students could be categorized into four distinct groups based on their unique patterns of Internet access. Notably, the study successfully identified 350 students who exhibited remarkably high levels of network usage, allowing for a comprehensive examination of how their academic performance and other achievements were influenced. This research, driven by the power of data, provides a practical and tangible demonstration of effectively utilizing data science in student affairs management. It presents compelling evidence that supports the accurate and scientific development of student affairs management practices by providing robust data support and generating invaluable insights.