Europeana is an aggregating digital library that aims at referencing all the resources digitized by museums and cultural heritage institutions in Europe. Europeana resources are described according to a Europeana Data Model (EDM) and their descriptions are compiled in a catalog that, at the time of writing, holds almost 50 million entries. This catalog can be searched from the Europeana website or by using an Application Programming Interface (API).
The Learning Resource Exchange (LRE) is an aggregation service offered by European Ministries of Education that are members of European Schoolnet. The LRE enables European primary and secondary school teachers to find quality Open Educational Resources (OERs) from many different countries and providers. The principle upon which the LRE is based is similar to the way Europeana operates. The LRE aggregates descriptions (i.e., metadata) of OERs and compiles them into a searchable catalog service. E-learning platforms (e.g., Virtual Learning Environments, Educational Portals) connect to the service using LRE Application Programming Interfaces (APIs). These platforms are the primary customers of the LRE. Although the LRE is open to any educational partners, it is used mainly by Ministries of Education (e.g., Flemish Community of Belgium, Sweden, Portugal, Lithuania, and Czech Republic).
Given the specificity of its audience, the LRE catalog does not aim at referencing all the OERs available in Europe and beyond but rather resources that travel well across national, cultural and linguistic boundaries so that they can usefully complement the national resources already offered by Ministries of Education.
In the framework of the Europeana Digital Service Infrastructure project funded by the Connecting Europe Facility (CEF) Trans-European Telecommunications Networks Work Programme 2014 of the European Commission, we worked on putting in place mechanisms to add a selection of relevant Europeana resources to the LRE catalog.
Adding Europeana resources to the LRE catalog involved the following steps:
- Resource identification: Europeana contains millions of digital resources, many of which hold little interest for educators teaching at the K-12 levels in Europe therefore it was necessary to identify relevant subsets for the LRE (where the ultimate users are K-12 teachers). Thus the initial effort focused on identifying a subset of Europeana collections that has both an educational value and the potential to “travel well”;
- Metadata acquisition: Once relevant resources have been identified, a second step consists of preparing queries necessary to obtain the metadata records describing these resources by calling the Europeana API;
- Metadata mapping and vocabulary crosswalks: In the LRE catalog, OERs are described according to the LRE Metadata Application Profile, which significantly differs from the Europeana Data Model and its controlled vocabularies. Therefore, in order to import metadata records into the LRE, it was necessary to map them from the Europeana Data Model (EDM) to the LRE Metadata Application Profile (LREMAP);
- Metadata enrichment: In order to increase their retrieval and reuse, the metadata descriptions of resources considered as especially interesting from a pedagogical and travel-well standpoint had to be further enriched with educational metadata; and
- Automation: Finally, the whole process was automated so that it can be repeated on a regular basis (i.e., at least weekly). This guarantees that relevant resources newly added to Europeana are automatically included in the LRE catalog.
In order to implement this approach, we applied the logic depicted in the Figure below.
In a preparatory phase, interesting records were first identified using the query system of Europeana. Then the queries were stored in a registry together with educational and other relevant metadata. In other words, a LREMAP metadata record was created for each query for describing the commonalities of all the resources returned by that query.
Once the registry was ready, the queries were launched (Metadata Acquisition) and the resulting EDM records were transformed into LREMAP records using a LRE to Europeana mapping and its associated vocabulary crosswalks (EDM to LREMAP Mapping).
Finally, the metadata records obtained by querying Europeana were enriched with the record used to describe the original query providing complete records (Metadata Enrichment) that were then injected into the LRE catalog.
This process is repeated weekly in order to bring in new relevant records.
Implementing and automating all of the above workflow was greatly simplified by using Tuktu, a new open source big data analytics platform that focuses on ease of use. Tuktu has now become our main platform and we’ll describe it in greater detail in our next posts.