Analysing the Quality of User Stories in Open Source Projects

Abstract

Agile Software Development has become highly popular over the last two decades. To-gether with the increase of popularity amount of scientific research on this topic has also increased. This research concentrates on one component from Agile Software Development: User Stories. In a recent paper quality framework for User Stories was proposed together with a tool implementing the framework. The proposed framework and tool make it possible to analyse the quality of User Stories from a new viewpoint. The main goal of this master thesis is to find if it is possible to forecast the quality of User Stories for monitoring pur-poses. Another goal is to find what is the relationship between User Story quality and other aspects of software development. Aspects studied in this thesis are the number of bugs, rework, and delays. The conducted analysis considers 10 open source agile software projects from which 8 are included in the analysis. Results of analysing the User Stories of these 8 projects revealed that it is possible to forecast the quality of User Stories. On average devel-oped forecast models were 94.5% accurate when producing predictions for the next 15 sprints. Regarding User Stories quality relationship to other software development aspects, several interesting patterns were found. In short, a decrease in User Story quality scores reflected in the increase in the number of bugs after approximately 3 to 4 weeks. The same happened for rework but the relationship took longer to reveal (approximately 4 to 6 weeks). Correlation of delays was more spread out meaning that the quality of User Stories was leading the interaction for a longer time.