Application of data anonymization in Learning Analytics

Abstract

Thanks to the proliferation of academic services on the Web and the opening of educational content, today, students can access a large number of free learning resources, and interact with value-added services. In this context, Learning Analytics can be carried out on a large scale thanks to the proliferation of open practices that promote the sharing of datasets. However, the opening or sharing of data managed through platforms and educational services, without considering the protection of users' sensitive data, could cause some privacy issues. Data anonymization is a strategy that should be adopted during lifecycle of data processing to reduce security risks. In this research, we try to characterize how much and how the anonymization techniques have been used in learning analytics proposals. From an initial exploration made in the Scopus database, we found that less than 6% of the papers focused on LA have also covered the privacy issue. Finally, through a specific case, we applied data anonymization and learning analytics to demonstrate that both technique can be integrated, in a reliably and effectively way, to support decision making in educational institutions.