The learning analytics architectural lifecycle

Less than a decade old, Learning Analytics is a new field in which research and practice are fragmented, new data sources are becoming widely available, whose value to education has yet to be thoroughly researched, and there is much duplication and fragmentation of the fields effort. To provide a safe path for implementing scaled LA services based on novel data sources, I defined a lifecycle with validated stepping stones. The lifecycle is split into three interconnected parts: i) strengthening and monitoring the current infrastructure, ii) understanding the context of new services, and iii) inclusive community practices. I highlighted the need for inclusive practices, and showed that there are many future research directions possible, and argued for an emphasis on community practice with the inclusion of all stakeholders. Through these means, we decrease the risks associated with deploying comprehensive services.
The outcomes included an international discussion on curation processes around capturing digital traces and associated software artefacts to support standardisation. An experimental open Synthetic Data Generator that has tested the stability of national experimental LA infrastructure and discovered bugs and performance issues. Validation of a novel data source, 3 million job advertisements for their predictive value in explaining longitudinal gender segregation. The application of job adverts as a source of recommendations within Eportfolios to improve the usability for lifelong learners and their teachers/mentors. The peer review and publication of design patterns that enable the large-scale deployment of LA services that support the lifelong learner. The organisation of many International hackathon events.