Learning Resource Metadata/Schema in action: Maths Apps Index

One of the nice things working for CETIS is when organisations like the Association of Learning Technologists (ALT) approach you to contribute to their projects it’s often an opportunity to field test cutting edge innovation. This was the case when ALT recruited me* for their Maths Apps Index project which is part of the maths4us initiative being co-ordinated by NIACE.

*I only work 0.8FTE for CETIS giving me room to explore other projects

Maths App Index is designed as a community review site for maths related resources. The idea builds on the Jisc funded ‘Community-led Evaluation and Dissemination of Support Resources – Pilot’, which I was also involved with. One of the recommendations made as part of this pilot was the better display and indexation resource reviews. When ALT asked me for guidance on the latest project and having kept abreast with the work my colleague Phil Barker was doing with new learning resource metadata standards my immediate response was to use the Learning Resource Metadata Initiative (LRMI) properties being proposed for schema.org.

LRMI/schema.org

For those unfamiliar with LRMI/schema.org below is a short briefing note I prepared as part of the project:

There is a general trend in webpages away from hidden metadata (keywords, descriptions contained in the header of a page) towards structured markup. This is in part a move by search sites to prevent the manipulation of search ranking by using hidden metadata. The solution has been to move towards combining human-oriented resource description and machine readable metadata.

An example of this is used in the Creative Commons embeds information about licenses in webpages. Using their ‘license chooser’ tool it generates extra HTML code for you to include in your distributed work. As well as the human readable icon and/or text the machine readable markup includes the rel="license" attribute shown in Figure 1.

clip_image002
Figure 1 Example of RDFa markup used in Creative Commons license

The inclusion on rel="license" allows search engines to identify that a resource might be released under a specific license, this information being used as a means to facet search results.

Schema.org

A development in this area of particular significance is schema.org, an initiative involving Google, Yahoo, Yandex and MS Bing that aims to:

"… improve the web by creating a structured data markup schema supported by major search engines. On-page markup helps search engines understand the information on web pages and provide richer search results." (schema.org, 2013)

There are two aspects to schema.org; a syntax for encoding parts of a page to identify additional metadata, and a shared schema of item types and their properties to make it easier for search engines to consistently index information.

Learning Resource Metadata Initiative (LRMI)

The Learning Resource Metadata Initiative (LRMI) is working to extend the the controlled vocabulary used in describing educational resources which is compatible with schema.org and other systems. This will mean search engines will be able to understand the information on web pages describing learning resources and make it easier for users to find them.

Table 1 is an extract from the draft LRMI Specification version 1.0[1] and describes the metadata that could be used to describe a learning resource. Within schema.org/LRMI all properties are optional.

Table 1 LRMI Specification version 1.0

Property

Description

educationalAlignment

An alignment to an established educational framework.

educationalUse

The purpose of the work in the context of education.

● Ex: "assignment"

● Ex: “group work”

intendedEndUserRole

The individual or group for which the work in question was produced.

● Ex: "student"

● Ex: "teacher"

interactivityType

The predominate mode of learning supported by the learning resource. Acceptable values are active, expositive, or mixed.

● Ex: "active"

● Ex: "mixed"

isBasedOnUrl

A resource that was used in the creation of this resource. This term can be repeated for multiple sources.

● Ex: "http://example.com/great-multiplication-intro.html"

learningResourceType

The predominate type or kind characterizing the learning resource.

● Ex: "presentation"

● Ex: "handout"

timeRequired

Approximate or typical time it takes to work with or through this learning resource for the typical intended target audience.

● Ex: "P30M"

● Ex: "P1H25M"

typicalAgeRange

The typical range of ages the content's intendedEndUser.

● Ex: "7-9"

● Ex: "18-"

useRightsUrl

The URL where the owner specifies permissions for using the resource.

● Ex: "http://creativecommons.org/licenses/by/3.0/"

● Ex: "http://publisher.com/content-use-description"


[1] http://wiki.creativecommons.org/LRMI/Properties

It’s worth noting that these were the draft specification and since then intendedEndUserRole has become educationalRole and, as noted by Phil useRightsUrl hasn’t currently made the cut.

In action

Because we thought it was unrealistic for a reviewer to supply data like educationalAlignment for the site we opted for a subset of the LRMI markup. As the review site is a WordPress installation we use a user submitted blog post for each review using the TDO Mini Forms plugin to capture the additional metadata. This is then rendered in modification of the Sampression Lite theme. Below an example of how LRMI is included within this review.

example LRMI/schema.org 

Where to next

When we plug the page into Google’s Structured Data Testing Tool we can see Google is detecting all our lovely metadata. Within a Google Custom Search Engine (CSE) we can even use this to filter a search. For example here are all the reviews with educationalUse set to independent learning. The problem however is Google CSE doesn’t currently provide any tools to all easy faceting of a custom search. Within the Maths App Index site we do have filtering for LRMI properties but currently it’s not integrated with search e.g. independent learning reviews. So with little perceived benefit why are we capturing this data now. It’s primarily about future proofing. I for one wouldn’t like to go back over thousands of reviews adding the metadata.