Mining GitHub For Monitoring And Evaluating Team Collaboration Process Among OSS Projects

With the development of the open source community,more and more people are participating in the social programming of GitHub.In order to achieve teamwork better,computer-related professional teachers also encourage students to carry out project development collaboration on GitHub.For the curriculum practice tasks that students collaborate on on GitHub,how teachers measure the contributions of members within the project team becomes a problem.The traditional scoring strategy mainly relies on human subjective evaluation and does not fully utilize the data information of the project team development process.This article makes full use of the large amount of data accumulated by software development activities and conducts research from the following four sections:(1)GitHub project team member contribution metrics analysis,helps teachers understand the individual contribution of the team.A project team member contribution index model is constructed,a set of calculation methods for contribution indicators is proposed,data collection scope is determined according to the contribution index model,and the individual evaluation of th e members within the project is completed through data processing calculation and visualized.(2)GitHub project team collaboration analysis,enables teachers to master the internal cooperation and communication within the team.A collaborative relationship network graph model is constructed,and the visualization of collaborative relationships is completed after data collection and processing.The degree of collaboration and the balance of collaboration relationships are defined to represent the degree of close coordination between the members of the project team.The measurement calculations complete the overall evaluation of the collaboration relationship.(3)GitHub project overall promotion characteristics analysis,helps teachers to obtain the progress information of the project development process from a macro perspective.Work pattern(abbr.WP)is defined,and the three attribute features needed in the work pattern mining are obtained through data processing,that is,add code trend,delete code trend and commits trend during the overall progress of the project.The atomic working mode unit niWT is obtained by decomposition,and the method of mining the working mode is given,and the seven working modes obtained by mining are analyzed and evaluated.The method for evaluating the ov erall progress of the project is proposed.(4)GitHub project team members personalized work habit analysis,makes teachers more in-depth understanding of students' development habits.Based on the LDA method to process code texts,a personalized analysis of the software function development involved in the project members is completed.In addition,the personalized work habits of the project members are analyzed from the aspects of working hours and GitHub activity types and are visualized finally.This paper proposes a comprehensive evaluation system for GitHub project team members through the mining of GitHub project team data and comprehensive evaluation of the project team(1),(2)and(3).The project team members can measure the contribution of project members more fully and help teachers to reduce the amount of assessment tasks.In addition,the work habits of the members of the project are analyzed to help teachers understand the students' development habits.Finally,the system design and implementation of the research results are presented in Web form.