The Implementation Process of Learning Analytics

With the popularity takeoff of the learning analytics area during the last decade, numerous research studies have emerged and public opinion has echoed this trend as well. However, the fact is that the impact the field has had in practice has been quite limited, and there has been little transfer to educational institutions. One of the possible causes is the high complexity of the field, and that there are no clear implementation processes; therefore, in this work, we propose a pragmatic implementation process of learning analytics in five stages: 1) learning environments, 2) raw data capture, 3) data tidying and feature engineering, 4) analysis and modelling and 5) educational application. In addition, we also review a series of transverse factors that affect this implementation, like technology, learning sciences, privacy, institutions, and educational policies. The detailed process can be helpful for researchers, educational data analysts, teachers and educational institutions that are looking to start working in this area. Achieving the true potential of learning analytics will require close collaboration and conversation between all the actors involved in their development, which might eventually lead to the desired systematic and productive implementation.