More Software Analytics Patterns: Broad-Spectrum Diagnostic and Embedded Improvements

TL;DR

This paper extends a set of software analytics patterns with two new patterns, Broad-Spectrum Diagnostic and Embedded Improvements, based on case studies and literature validation, to enhance decision-making in software development.

Contribution

It introduces two novel patterns for software analytics adoption, expanding the pattern language to support broader and embedded decision-making improvements.

Findings

Patterns validated through literature search

Case studies informed pattern development

Enhanced decision-making guidance in software analytics

Abstract

Software analytics is a data-driven approach to decision making, which allows software practitioners to leverage valuable insights from data about software to achieve higher development process productivity and improve different aspects of software quality. In previous work, a set of patterns for adopting a lean software analytics process was identified through a literature review. This paper presents two patterns to add to the original set, forming a pattern language for adopting software analytics practices that aims to inform decision-making activities of software practitioners. The writing of these two patterns was informed by the solutions employed in the context of two case studies on software analytics practices, and the patterns were further validated by searching for their occurrence in the literature. The pattern Broad-Spectrum Diagnostic proposes to conduct more broad analysis based on common metrics when the team does not have the expertise to understand the kind of problems that software analytics can help to solve; and the pattern Embedded Improvements suggests adding improvement tasks as part of other routine activities.