Interoperable Data Collection
We would like to announce the formation of SPLICE Working Group on Interoperable Data Collection. As a focus on standards and data interoperability gains momentum, we invite SPLICE community to take an active part in forming its agenda as well as participate in its meetings and other activities.
The focus of the group is collecting and representing heterogenous student activity traces in an industry-standard interoperable fashion that would enable collection, transfer, archival storage, sharing, and issue-free posthoc analyses. The workgroup would engage in a set of activities that target the creation of processes and deliverables addressing learning data interoperability in observance of de-facto industry standards and wide-spread solutions.
Tentative Technologies and Tools
To achieve the stated goals, the following suite of technologies and solutions is considered.
- Callisto as a hosted Learner Record Store working with xAPI data format and supporting managed access to the data.
- Amazon Web Services as a cloud platform for hosting the Learner Record Store.
- Docker as the technology for supporting containerized development of the LRS and data stream solutions.
- Postman (optionally or complementary Apache JMeter) as the technology for testing the developed solutions via managed data-light request collections.
A working draft of activities to be carried out to complete this project is as follows.
- Hold regular meetings of the workgroup.
- Align data stream representations within the SPLICE universe to the xAPI/Caliper format. This activity focuses on creating a set of recommendations for serializing heterogeneous source data.
- Create a working example of a containerized (using Docker) and deployed (using AWS) instances of xAPI LRS.
- Create an example suite of request collections for testing the correctness of operation and integrity of the containerized/deployed xAPI LRS.
- Author and publish requirements for importing previously collected heterogeneous data streams and executing multiple sample jobs as a proof-of-concept.
All interested in the workgroup's activities please, contact the organizer, Michael Yudelson