Event location: Prestonfield room, John McIntyre Conference Centre
This two-day pre-conference hackathon on April 25-26, 2016, brings together two organisations leading the way worldwide in developing open architectures for learning analytics. Jisc and Apereo will put the growing ecosystem of learning analytics products such as learning records stores, learning analytics processors, dashboards, consent systems and student apps through their paces with synthetic big data from learning management systems, student record systems and other sources.
Testing the interoperability of the various tools, and integrating new data sources, predictive models and third party products will help to reassure institutions that they are not going to be locked into proprietary learning analytics systems and can select the best products for their needs. xAPI, an emerging standard for the exchange of learning records, will be the basis of efforts to develop new profiles (aka “recipes”) and applications for learning analytics.
Getting some of the leading technical minds in the field together in a room for two days aims to help expand the market for experienced vendors as well as innovative new entrants. Jisc is inviting the developers of the various components of its open learning analytics architecture to participate in the workshop, and Apereo will input the growing infrastructure of its learning analytics initiative to the event. Vendors and developers of other commercial and open source learning analytics related products are encouraged to sign up for the hackathon.
Day One program
Day One will aim to generate a common understanding about how data portability can be enabled between the many different stakeholders involved in the xAPI community. We will start with a consideration of how learning activities and high-level dashboards should drive the development of xAPI recipes. Concepts such as discourse analytics, metacognition, the achievement of competency, and personalisation will all be considered from within an xAPI based system – will the current profiles enable or restrict these high-level use cases?
We will work to modify and unify the recipes currently used by Jisc, considering ways in which they might be enhanced to facilitate fast and useful learning analytics functionality. Making reference to other extant recipes already produced by the community, we will identify a number of important use cases for Learning Analytics systems built using an xAPI data format. These will then drive the development of common vocabularies, sophisticated profiles and a more unified treatment of learning contexts. Meanwhile, an elementary synthetic data generator will be implemented based on participants’ recommendations to develop a performance test suite.
Day Two program
Day Two will continue from Day One, including the new xAPI recipes and the synthetic data generated, will be used to inform suggestions for data structures and to bootstrap high-level analytics solutions and generic dashboards. Using Jisc’s open architecture for learning analytics and a reference university implementation, the following activities are anticipated:
Seeing if the new xAPI references developed on Day 1 can be implemented in practice
Developing new visualisations based on the APIs (and also feedback on how the APIs work for this purpose)
Using the data to start looking at new predictive models
Creating new apps and applications doing for novel things with the APIs
By the end of the hackathon, we aim to have built understanding and achieved greater consensus among participants around the key APIs for learner data, learning data and predictive models. We should see some new systems built and some existing ones refined to achieve greater interoperability between the various components of the rapidly developing learning analytics ecosystem.
Josh Baron, Marist College
Alan Berg, University of Amsterdam
Ian Dolphin, Apereo Foundation
Kirsty Kitto, Queensland University of Technology
Niall Sclater, Jisc
Michael Webb, Jisc
Sandeep Jayaprakash, Marist College