Experiences tracking agile projects: an empirical study

Abstract

In this article, we gather results from several projects we conducted recently that use some kind of agile method. We analyze both academic and governmental software development projects, some of them using agile methods since the beginning and others in which agile methods were introduced afterwards. Our main goals are to classify the different projects, and to analyze the collected data and discover which metrics are best suited to support tracking an agile project. We use both quantitative and qualitative methods, obtaining data from the source code, from the code repository, and from the feedback received from surveys and interviews held with the team members. We use various kinds of metrics such as lines of code, number of tests, cyclomatic complexity, number of commits, as well as combinations of these. In this article, we describe in detail the projects, the metrics, the obtained results, and their analysis from our main goals standpoint, providing guidelines for the use of metrics to track an agile software development project.