The Design And Implementation Of Visual Analytics System For Group Discussion Teaching

Group discussion refers to how students communicate and explore the designated problems in groups under the guidance of teachers.As a critical student-centered instruction method,group discussion can stimulate students’ learning initiative,which is a hot issue in the current educational reform study.However,in the practice of teaching,it is difficult for teachers to accurately get the precise situation of group discussion,which makes this instruction form challenging to achieve the ideal outcome.The specific manifestations are as follows: 1)browsing the lengthy audio and video recordings of the discussion is quite time-consuming,which makes it difficult for teachers to summarize the progress of the agenda discussed quickly;2)The interactive relationship between students can not be obtained directly from the original data and will change dynamically with time.It needs to be analyzed and evaluated in combination with the specific context,which challenges teachers to understand the interaction between students;3)Students’ performance is affected by many factors.It is difficult for teachers to reasonably evaluate students’ performance and analyze the reasons behind their performance.Based on the above challenges,this study proposes a visual analysis system for group discussion,integrating teachers’ experience and knowledge into the analysis procedure.It helps deeply explore and analyze the multimodal data,including video,audio,text,and metadata,to support teachers in better understanding and following the development of group discussion.The main work of this paper is as follows:(1)Determining the analysis tasks and design requirements according to the interview results with education experts and the characteristics of group discussion data,which guide the following visual design.(2)Proposing a data processing flow and a series of visual designs based on the analysis task and design requirements.Specifically,an agenda segmentation algorithm based on the response relation tree and a visual design based on the enhanced calendar graph is proposed to identify and visualize the progress of agendas respectively;Five important students’ performance features are identified and extracted,and the corresponding visualization is designed according to the focus of experts;A layout optimization algorithm is proposed to reduce the visual confusion of the sequential node-link diagram,and based on this algorithm,a novel interactive relationship view is proposed to show and the evolution of the interactive relationship between students over time.(3)Combined with the proposed view and various interactive technologies,propose a visual analysis system,allowing teachers to explore the group discussion from three different aspects.The availability and effectiveness of the system are proved by case analysis and expert interviews.