‍

Towards Learning Analytics Scalability:

Enriching Context Descriptions Across Data Sources

‍Abstract

Following advances in technology, digital tools and systems related to education are increasingly being adopted by educational institutions, including learning management systems, student information systems, exam systems, and adaptive learning tools. Such tools and systems generate and store vast amounts of data pertaining to learning. Learning Analytics (LA) is a field concerned with data collection, analysis and reporting regarding learners and their context, aimed at providing insights and improvements related to learning and learning environments. Data integration, the collection and combination of data originating from multiple data sources, can contribute to more precise and useful LA results based on increased data diversity and scale. However, data originating from different sources are often stored in different formats, with varying levels of structure, using different storage technologies, resulting in silos where data are prevented from being integrated and analyzed across sources. As such, data integration has been identified as a key challenge in LA. Data integration is an important aspect of LA scalability, and is tightly linked to interoperability. To support data integration, the use of a learning activity data standards such as xAPI enables the representation of learner data in the most basic form of actor verb object, i.e., a learner interacting with a learning object (in a learning environment), where further syntactic structures are included that allow for the registration of learning context data.

This research aims to gain insight into, and address challenges for, LA data integration,i nteroperability, and consequently scalability. More specifically, it is focused on identifying gaps in xAPI expressibility and providing solutions to these gaps.

Following a design science research methodology, the research carried out a systematic review on the state of the art related to LA data integration, and a case study. The case study used the the Activity Data for Assessment and Adaptivity (AVT) project that utilizes xAPI for describing K-12 learner data from multiple data sources as the case, and comprised two empirical studies which included interviews and user testing with stakeholders.

Bio

Jeanette Samuelsen (born 1984) has a Masters degree in Information Science, University of Bergen. Currently she is working as a researchers at the Centre for the Science of Learning & Technology (SLATE), University of Bergen.

‍

Trial Lecture and PhD Defense

Jeanette Samuelsen's trial lecture takes place at Sydneshaugen skole, Auditorium B, at 09:15 – 10:00 on Friday, 18 August. The topic of the lecture: A systematic comparison of xAPI and Caliper: Extending the state of the art. The PhD defense is at 10:30 – 14:30, and will take place at the same venue.

The first opponent is Professor Hiroaki Ogata, University of Kyoto, Japan. The second opponent is Senior Researcher María Jesús Rodríguez Triana from Tallinn University, Estonia.

Professor Barbara Wasson, Director at SLATE, is Jeanette's main PhD supervisor and Professor Weiqin Chen, OsloMet & SLATE is the co-supervisor.

‍

Jeanette’s trial lecture and PhD defense are open to everybody, and all are welcome to attend! There will be a reception at SLATE (Christiesgate 12, 2nd floor) after the disputas.

‍