Measures of Software Development Leading to Applications Enhancements

Abstract

This study analyses measures and patterns from data drawn from application communications and updates from development teams, to assess the value of data-driven prioritisation of software enhancements, and to develop a framework to facilitate decision-making. The data source used is GitHub and datasets include software user interactions that influence the software development lifecycle of projects such as feature requests, bug reports and issues logged and their resolutions. The dataset also includes software developer activities against a software projects such as code commits, cloning of repositories, peer review activities relating to code changes and releases of changes. Multivariate time series data analysis methods are applied to identify which factors influence software changes. The results are used in this study to create models and frameworks for decision-making based on available data.

Businesses have encountered rapid changes in customer demands for content and quality of software delivered in short time frames resulting in software development teams adopting ways of working and technologies to release software as Minimal Viable Products with continual delivery of enhancements. This presents an opportunity to use metrics to prioritise product enhancements to continuously deliver valuable software. However, there exist opportunities for changes to be made that do not provide value and this study seeks to establish value can be ascribed.