An Updated Analysis of Learning Resource Metadata Usage on the Web

Ratan Sebastian, Anett Hoppe

Abstract: Tools for discovering learning resources become increasingly important as more and more educational offerings move online. Improvements in the retrieval or recommendation of these resources often rely on the availability of metadata. For example, it has been demonstrated that showing teachers educational metadata alongside the search result could improve search outcomes. However, this relies on relevant educational metadata being embedded in web pages using formats such as JSON-LD or MicroData. For learning resources, the Learning Resource Metadata Initiative (LRMI) ontology defines classes and properties to express such embedded educational metadata. Previous studies have assessed its adoption, quality, and conformance to the ontology prior to an LRMI version update in 2020. This contribution updates prior studies with respect to adoption following the version change. We then focus on mining usage patterns of LRMI properties for the benefit of application developers who would like to leverage this resource. We also expand our analysis beyond just resources that use LRMI by training a text classifier to identify educational web pages. As a result, we are able to present what we believe to be the broadest and most recent examination of usage patterns and adoption of learning resource metadata on the web.