Abstract:

In 2025, Northumbria University, UK, embarked on an ambitious programme to explore the opportunities of Microsoft Fabric to deliver analytics insights on student engagement data. This builds on a noted sector reputation for innovation in learning analytics (Foster and Francis, 2020) and student mental health analytics (Foster, 2025).

As Microsoft Fabric was only unveiled by Microsoft in 2023, real world examples are still rare. As such we secured funding from Microsoft for a short development sprint to deliver a proof of concept (POC) which took source data from the university management information systems in order to predict students' success outcomes.

The focus of the POC was different from other analytics projects which typically focus on the quality of the prediction, the optimal modelling techniques and feature exploration of data variables. Rather, this POC aimed to highlight the necessary operational preparation required to move away from a commercial Edtech platform, in situ for 9 years, to an in-house built system which will be built in partnership with a Microsoft tech partner and maintained by the university’s own development team.
The POC identified several useful findings such as the upskilling required on programming language, the licensing requirements and governance structure of the power platform space and the suitability of datasets to machine learning within the Fabric space (including the AutoML functionality).

This presentation will offer insights into how universities may seek to leverage their existing enterprise partnerships with Microsoft or similar providers and what may be required to opt for or transition to such as approach. Whether from a technical or student experience background, it will offer valuable and practical insights that will inform future business cases for exploring analytics projects.

This is the first known deployment of Microsoft Fabric in the Higher Education sector.

Keywords:

Analytics, technology development, prediction.