Abstract:

This paper is motivated by the lack of research dealing with modeling, building, using and maintaining infrastructure for learning analytics (LA) at HEIs, especially in blended learning environments.

The main aim of this research is to provide a conceptual model of infrastructure supporting LA in blended learning environment.

The research is organized around two research questions:
- What are the existing models of LA infrastructure?
- How to build a model to utilize LA in a blended teaching environment?

To answer the research questions authors performed systematic literature review and proposed conceptual infrastructure model for LA in blended learning environment consisting of three layers and following components:
1. Data sources,
2. Data storage and analysis, and
3. Visualization and action.

This layered approach is expected to enable IT specialists to combine various solutions in each of the layers and its components, enabling them to create flexible infrastructure, capable of changing according to future needs while reducing the overall implementation and maintenance costs.

Although LA is a very young research field, developing very quickly with an exponential growth in the number of papers, its practical use HEIs is extremely low. The literature review showed that the number of papers describing the field, its possibilities and challenges is rich and offers different perspectives for further research. Still, the number of papers describing and discussing various options for implementation of HEI LA infrastructure solutions is very limited. As Salonen et al. (2021) conclude, LA provides educators with a means of being informed about the learning process. This helps educators to enhance their teaching practices, assess the impact of teacher interventions, and improve the structure of study units.

Further, the HE sector has always been a data-rich sector, and universities generate and use enormous volumes of data each day. However, the sector has not yet capitalized on the enormous opportunities presented by the data revolution, and is lagging behind other sectors in this area. Moreover, unless institutions and university staff are data-capable and equipped with the resources and skills to manage data well, HE will not be able to catch up and students will miss out on many potential learning and support benefits.

As LA data provides a snapshot of how engaged students are and how well they are performing, this could be considered a useful indication of where excellent teaching is taking place. The EC recommends that institutions should be encouraged to use the information from LA systems to identify and foster excellent teaching within their institutions. The list of recommendations is provided in the report The Potential of Data and Analytics in Higher Education (2016).

As stated above, the proposed infrastructure model consists of three layers in order to enable IT specialists to create flexible infrastructure while reducing the overall implementation and maintenance costs. Data sources, as a precondition for implementation of LA, need to be diverse, reliable and it needs to support collection of different types of data such as grades, sociodemographic data, background data, data on learners’ behavior in academic environment (student activities, mobility, internship).

Therefore, it is expected that this conceptual model will help HEI management and provide a foundation for understanding how they can make LA adoption decisions.

Keywords:

Learning analytics, higher education institutions, infrastructure, conceptual model, blended learning.