Faceted search

Faceted search is a technique used in search interfaces, particularly within dataset catalogues like those using the Dataset Catalogue Vocabulary (DCAT) standard.

 
 
It allows users to filter and refine search results using multiple attributes or categories, known as facets. These facets are usually derived from metadata properties associated with the datasets. Faceted search allows users to efficiently narrow down large volumes of datasets based on specific criteria. Each facet represents a metadata element that describes the datasets, and users can select one or more facets to refine their search results. For example facets might include: keywords, publisher, themes or categories, temporal coverage, geographic coverage, data formats, licences, ... When a user selects a facet, the catalogue dynamically filters the list of available datasets to only those that match the selected criteria.

To make faceted search operational, DCAT-AP provides already a bulk of key metadata elements: Keywords (dct:keyword), Publisher (dct:publisher), Theme (dcat:theme) linked to a controlled vocabulary helping users to find datasets based on subject areas like health, Temporal Coverage (dct:temporal), Geographic Coverage (dct:spatial), Format (dcat:mediaType or dct:format),... Other elements can be used to enhance the faceted search capabilities depending on the specific use case. HealthDCAT-AP has introduced a number of new facets to improve the search user experience.
Figure 4: Optimised Filters for HealthDCAT-AP Dataset Catalogues: Dynamic Display Based on Personal Data Inclusion
For instance, a mandatory controlled vocabulary denoting health data within the scope of the Commission Regulation (ref: Art.33 - 15 top health categories) MUST be used in healthdcatap:healthCategory to enable the 'Category' facet. Additional facets specific to HealthDCAT-AP include:

  • Health Data Access Bodies,
  • Publisher Type,
  • Trusted Data Holder,
  • Personal Data Classes,
  • Age Ranges,
  • Population Size,
  • Quality Label
  • and Coding Values.

These new properties introduced in HealthDCAT-AP significantly enhance the description of health datasets and are designed to improve the data search experience.
Implementation consideration
To implement effective faceted search, metadata must be sufficiently detailed. In this context, the implementing act that will define the minimum metadata elements is essential for the future performance of health dataset catalogues. (i.e., minimum metadata elements for Art.51 and Art.80)
EHDS Regulation Article 77) Dataset description and dataset catalogue
4. ... The Commission shall, by means of implementing acts, set out the minimum (x) elements health data holders are to provide for datasets and the characteristics of those elements. Those implementing acts shall be adopted in accordance with the examination procedure referred to in Article 98(2).
In addition, the search interface must be designed to present these facets clearly and intuitively to users. It is important to consider UX improvement in two ways: prioritising the facets and dynamically adapting the list of filters.

- Prioritising the facets: Not all facets have the same level of relevance, and the pilot use cases have shown that data queries can have both commonalities and specificities (e.g., searching for a theme or a specific dataset format). The filter interface should be designed based on these insights.

- Dynamically adapting the list of filters: A health dataset catalogue governed by Article 33 encompasses both personal and non-personal electronic health data. Several metadata elements introduced by Health DCAT-AP, which function as facets, are applicable only to specific types of data (e.g., either personal or non-personal). As a result, the filters can be dynamically adjusted based on the user's search criteria, ensuring that users are presented with the most relevant options. (See Fig. 4: 'Optimised Filters for HealthDCAT-AP Dataset Catalogues: Dynamic Display Based on Personal Data Inclusion')