Reducing integration complexity of learning analytics in eXtended reality through a modular and standards-based pipeline : design and evaluation of the OmiLAXR framework

RWTH Publications (RWTH Aachen)

Abstract

Despite the increasing prevalence of eXtended Reality (XR) in education, heterogeneous technology stacks and high integration costs make it difficult to collect and reuse XR learning data in a reliable and comparable manner. This dissertation investigates how the integration complexity of Learning Analytics (LA) in XR can be reduced without compromising reliability, interoperability, reproducibility, or comparability. It introduces Open and modular integration of Learning Analytics in eXtended Reality (OmiLAXR), a modular pipeline that separates the acquisition of XR signals and events (Data Event Layer (What, When) (DEL)) from their data-standardization, processing, and storage (Data Standard Layer (How, Where) (DSL)) across six phases. The phases provide plug-and-play integration and interface-based extensibility, with a strictly typed eXperience API (xAPI) statement construction serving as an example DSL.OmiLAXR was developed across three design iterations, following the Design-based Research (DBR) process, and refined through deployments in research and teaching contexts, as well as through stakeholder feedback. The resulting Research Software (RS) artifacts were assessed against the FAIR Principles for Research Software (FAIR4RS) and the CDUR procedure. The resulting RS artifacts were assessed using FAIR Principles for Research Software (FAIR4RS) and Citation, Dissemination, Use and Research (CDUR). A supplementary technical evaluation combined a controlled benchmark to quantify runtime behavior and xAPI statement throughput, while an Online Sustainability Evaluation (OSE)highlighted concrete gaps in documentation and operational maturity. The results show that OmiLAXR can be integrated with little effort. Under realistic conditions, OmiLAXR operates stably while maintaining xAPI statement throughput up to a conservative load level. Remaining deficits primarily concern ecosystem-level maturity. OmiLAXR has already been successfully used to collect data in dissertations, student classes and research projects. These applications demonstrate a certain maturity and reusability of this RS beyond its original development environment. Overall, this work aims to contribute to a reusable, standards-based integration infrastructure for XR-LA. The key innovation is to design XR-LA not as a one-off project-specific implementation, but as a modular pipeline with clear interfaces. This supports plug-and-play integration, extensibility, and interoperability, and facilitates empirical research through repeatable data collection and better comparability across projects.

Authors 1

1. Sergej Görzen corresponding Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen