Software metrics visualization

Abstract

Software engineering is an empirical field of study. To support managerial and technical decision-making, the engineer needs numerical measures closely connected with different software metrics. Visual representation of numerical data improves the effectiveness of human data processing and shows insights that humans may miss. This paper aims to provide a systematic review of the approaches for software metrics visualization and define the possible recommendation for their use. The study is based on the literature review of the papers from two text collections – IEEE Xplore and ACM Digital Library – and the scientometric database Scopus. After merging and filtering, the final set of publications contains 16 papers. Our study showed that there were the metrics used significantly more often; among them are lines-of-code, cyclomatic complexity, coupling, and cohesion. We were not able to identify such leaders for visualization means. Instead, there was a tendency to combine different metrics on one chart or dashboard to provide the whole process picture. Based on the results of empirical studies reported in the literature, we offered an analysis of simple charts’ properties and recommendations on their use for support decision-making in the software engineering process.