Software Analytics to Software Domains: A Systematic Literature Review

TL;DR

This systematic literature review analyzes the current state of Software Analytics, highlighting its focus on developer needs, limited artifact linking, and identifying gaps and future research directions in this emerging field.

Contribution

The paper provides a comprehensive review of SA research, filtering relevant studies, and identifying key gaps such as limited artifact linking and focus on developers.

Findings

Most SA research targets developers' needs

Limited linking between software artifacts in studies

Significant research gaps and future directions identified

Abstract

Software Analytics (SA) is a new branch of big data analytics that has recently emerged (2011). What distinguishes SA from direct software analysis is that it links data mined from many different software artifacts to obtain valuable insights. These insights are useful for the decision-making process throughout the different phases of the software lifecycle. Since SA is currently a hot and promising topic, we have conducted a systematic literature review, presented in this paper, to identify gaps in knowledge and open research areas in SA. Because many researchers are still confused about the true potential of SA, we had to filter out available research papers to obtain the most SA-relevant work for our review. This filtration yielded 19 studies out of 135. We have based our systematic review on four main factors: which software practitioners SA targets, which domains are covered by SA, which artifacts are extracted by SA, and whether these artifacts are linked or not. The results of our review have shown that much of the available SA research only serves the needs of developers. Also, much of the available research uses only one artifact which, in turn, means fewer links between artifacts and fewer insights. This shows that the available SA research work is still embryonic leaving plenty of room for future research in the SA field.