Research On Visual Analysis Of Multivariate And Hierarchical College Course Teaching Data

In-depth analysis of student performance in college courses and its influence factors is significant to curriculum arrangement optimization and teaching quality improvement.However,the student score data contains multiple relevant subjects and has characteristics like multivariate,multi-attribute,hierarchical and time-related.Traditional analysis tools and display methods are limited when exploring associations and anomalies in curriculum performance.Since information visualization and visual analysis technique could combine human experience,cognitive capacity and reasoning ability with automatic computer processing ability,and have advantages like intuitively acquiring patterns,quickly discovering anomalies and exploring causes through interactions,they apply to analysis of large complex data in various domain.Consequently,this thesis refines analysis tasks from multiple aspects of the courses,students and teachers.On this basis,student performance visual analysis techniques and a prototype system called SPVAS which consists of multiple coordinated views are designed according to the characteristics of the student score data.Firstly,to explore the temporal distribution,variation patterns of the student performance in different grades and semesters are discovered using heatmap matrix integrated with multi-attribute.Then,by extending interaction and demonstration ability of the parallel coordinates,multivariate statistical characteristics of the student performance and correlative subjects such as courses and teachers can be presented.Finally,influence factors of student performance and correlation among courses are revealed through novel layouts of combinations of arc diagram with both parallel coordinates and node-link diagram.With interaction techniques like cross filters and dynamic association in multiple views applying on the novel layouts,a complete visual exploration and analysis process of student performance is achieved.A case study based on a real dataset has demonstrated that SPVAS well supported cross analysis and coherent inference of regularities and anomalies contained in the student performance by starting from any aspects of course,student and teacher.The feedback from administrators has showed that this system can effectively reveal multiple correlated factors in the student performance and analyze anomalies from multiple aspects,meeting the decision-making needs of optimizing curriculum,allocation of teachers and course teaching.They consider SPVAS as a successful attempt which applies visual analysis technique in teaching management domain.