Learning Analytics is a rapidly growing research field and commercial , with potentially disruptive potential. While educationalresearchers have for many years used computational techniques toanalyse learner data, generate visualizations of learning dynamics,and build predictive models to test theories — for the first time, these techniques are becoming available to educators, learners and policy makers. Learning analytics promise is to transform educational research into a data-driven science, and educational institutions into organisations that make evidence-based decisions. However, critical debate is needed on the limits of computational modelling, the ethics of analytics, and the educational paradigms that learning analytics promote.