Data-Driven Software Engineering: A Systematic Literature Review

Over the past few years, emerging technologies have had a significant impact on the processes of software engineering (SE). Consequently, there has been a shift from a more experience-based approach to a data-driven decision-making approach. This shift to data-driven decision-making has resulted in more reliable and accurate decision-making, ultimately leading to more efficient and effective SE processes and a reduction in rework. Our study involved a comprehensive systematic literature review (SLR) examining the utilization of data-driven approaches in SE processes over the last decade. Our analysis of 34 primary studies revealed that data-driven approaches are most commonly utilized. After analyzing the primary studies, we found that data-driven methods are commonly employed in SE processes for software management and software testing. Researchers are delving into subfields of artificial intelligence, including machine learning and deep learning, to devise decision-making models for SE processes that have undergone extensive validation. We aim to provide valuable insights into the usage of data-driven approaches in SE by conducting a systematic mapping based on the studies that we have found.