The University of Science and Technology of China has developed a smart education analysis system capable of quantifying various aspects of classroom behavior and extracting comprehensive data on teaching and learning. Built on this system, the university has established a smart evaluation framework that provides robust support for ongoing improvements in teaching quality and educational management. The system efficiently captures classroom video signals and extracts information on classroom behavior by connecting with both the teaching data center and classroom video equipment. It employs technologies such as multi-view skeleton action recognition, multi-camera sensing, and intelligent data anonymization to convert behavioral observations into measurable data. The system then applies learning models and filtering adjustments to create specific metrics for intelligent analysis of classroom dynamics. Additionally, it can issue alerts regarding anomalies in educational processes, such as attendance issues, varying levels of engagement, and teacher-related concerns including tardiness, absences, and early departures. Timely detection and handling of such concerns promote refined educational management and scientific decision-making. Ultimately, the system presents educational data on a dashboard and generates analytical reports for individual schools and the entire university, providing visual data support for educational administrators to enhance educational strategies and improve the overall teaching quality