Software process anti-pattern detection in project data

There is a significant amount of guidance on Project Management (PM) including software development methodologies, best practices and anti-patterns (APs). There is, however, a lack of automated way of applying this knowledge by analyzing readily available data from tools aiding in software PM, such as Application Lifecycle Management (ALM) tools. We propose a method of detecting process and PM anti-patterns in project data which can be used to warn software development teams about a potential threat to the project, or to conduct more general studies on the impact of AP occurrence on project success and product quality. We previously published a concept for the data mining and analysis toolset distinct from other research approaches and related work. Based on this toolset, we devised a formalized basis for our detection method in the form of standardized AP description template and a model for pattern operationalization over project data extracted from ALM tools. The main contribution of this paper is the general method for AP operationalization taking the description template as a starting point, discussed together with its potential limitations. We performed an initial validation of the method on data from student projects, using an AP we encountered in practice called "Collective Procrastination" which we also describe in this paper together with its detailed formal operationalization.