Software quality is an important topic for software practitioners in order to guarantee how the system is built and performed. In last years, techniques related to data science have been utilized in software engineering field as sup-port for building mainly prediction models. These approaches focused on trying to minimize software problems during the development and performance of software, helping to make right decisions. This systematic literature review (SLR) aims at investigating the significant techniques of data sciences used in software processes, identifying their major impacts and problems/challenges of use. This review will be of interest for software practitioners concerned on software quality.