Abstract
Student employability is crucial for educational institutions as it is often used as a metric for their success. The job market landscape, however, more than ever dynamic, is evolving due to the globalization, automation, and recent advances in Artificial Intelligence. Identifying the significant factors affecting employability, as well as the requirements of the new job market can tremendously help all stakeholders. Knowing their weaknesses and strengths, students might better plan their career. Instructors can focus on more appropriate skill sets to meet the requirements of rapidly evolving labor markets. Program managers can anticipate and improve their curriculum to build new competencies, both for educating, training and reskilling current and future workers. All these combined efforts certainly can contribute to increasing employability. Data driven and machine learning techniques have been extensively used in various fields of educational data mining. More and more studies are investigating data mining techniques for the prediction of employability. Yet, these studies show a lot of variation, for instance, with respect to the data used, the methods adopted, or even the research questions posed. In this paper, we aim to depict a clear picture of the art, clarifying for each standard step of data mining process, the differences, and similarities of these studies, along with further suggestions. Thus, this survey provides a comprehensive roadmap, enabling the application of data mining for employability.