Reimagining the conceptualisation, design and delivery of Learning Analytics

Abstract

Learning analytics applications have the potential to bring value to educational environments by providing the ability to analyse and present data about students in ways that are meaningful and actionable for educators. Recent studies have highlighted the importance of involving educators in the design process for these systems to ensure that what is developed meets their needs. However, this can add significant time and resource investment to a development project and requires the involvement of a range of people with data science and developer skills. Across various industries, new approaches are being imagined and trialled to reduce reliance on technical staff and long development processes when designing data-based systems instead enabling users to create a story from data without needing programming skills. In this paper we review the innovations introduced by DIVE, a mixed initiative visualisation and analysis application developed by a team at the MIT Media Lab, and discuss how these can be adapted to the field of learning analytics. DIVE includes a range of features that can provide value to learning analytics applications such as automated dataset statistical analysis, visualisation recommendation, and data story generation. An analysis of these features informs a set of guidelines which will enable the reconceptualisation of co-design sessions, reducing the time required to design learning analytics applications and facilitating the automated generation of prototypes, and in some cases full dashboards and analytic applications. The paper concludes with a discussion of future enhancements that could be made to the functionality of DIVE, such as the need for a semantic layer on trace and assessment data, to enable the translation of data into meaningful insights on learning.