Categorization and Visualization of Issue Tickets to Support Understanding of Implemented Features in Software Development Projects

Abstract

Background: In most software projects, new members must comprehend the features already implemented since they are usually assigned during the project period. They often read software documents (e.g., flowcharts and data models), but such documents tend not to be updated after they are created. Herein we focus on tickets issued because they are created as a project evolves and include the latest information of the implemented features. Aim: The purpose of this paper is to clarify the way of helping new members understand the implemented features of a project by using tickets. Methodology: We propose a novel method to categorize tickets by clustering and visualizing the characteristics of each category via heatmapping and principal component analysis (PCA). Our method estimates the number of categories and categorizes issue tickets (tickets) automatically. Moreover, it has two visualizations. Ticket lifetime visualization shows the time series change to review tickets quickly, while ticket feature visualization shows the relationships among ticket categories and keywords of ticket categories using heatmapping and PCA. Results: To evaluate the effectiveness of our method, we implemented a case study. Specifically, we applied our method to an industrial software development project and interviewed the project members and external experts. Furthermore, we conducted an experiment to clarify the effectiveness of our method compared with a non-tool-assist method by letting subjects comprehend the target project, which is the same as that of the case study. These studies confirm our method supports experts’ and subjects’ comprehension of the project and its features by examining the ticket category lifetimes and keywords. Implication: Newcomers during project onboarding can utilize tickets to comprehend implemented features effectively if the tickets are appropriately structured and visualized. Conclusions: The original contribution of this paper is the proposal of the project feature comprehension method by visualizing the multi-dimensional nature of requirements in an organized and structured way based on available tickets and the result of its application to the industrial project.

Keywords

• software development projects
• software comprehension
• issue tickets
• text mining
• machine learning