FAIR Principles and Linked Open Data
From Linked Data and to Linked Open Data
Tim Barnes-Lee (2006), who created the World Wide Web and is the leading proponent of Linked (Open) Data, stated that the Semantic Web proposal is about creating links that are responsible for the semantics that allows data interpretation by machines. From this statement, following Ninin (2018) suggestion, we should distinguish two aspects to understand the Linked Data principles and the Linked Open Data movement, that are the availability of data on the Web and the practices of linking these data.
Since these principles relate directly to the Web platform, it is assumed that data are available for manipulation and use, according to the Open Data principles. In this principle, open data is data published under an open license. Berners-Lee (2006), however, rectifies his text by explaining that this is not always the case: data can be linked, but not necessarily open. In this sense, the Linked Data principles refer specifically to the procedures for linking data, open or not, through the systematic use of tools developed by the Semantic Web. Thus, by using the proven architecture of the World Wide Web, Linked Data aims to achieve interoperability at the widest scale.
The Linked Data paradigm postulates four rules:
- Use URIs as names for things
- Use HTTP URIs so that people can look up those names.
- When someone looks up a URI, provide useful information, using the standards (RDF, SPARQL)
- Include links to other URIs, so that they can discover more things.
The use of URI (first principle) provides a unique identification of a resource and its association with the Hypertext Transfer Protocol (HTTP) (second principle) allows it to be “findable” in the Web environment. This is because HTTP is a standard protocol for communication and location of documents, or resources on the Web. The description and the relationship with other resources (third principle) are made effective by applying the RDF model, while its retrieval and manipulation as raw data occurs through the use of the SPARQL protocol. The fourth principle, along with the third, is responsible for the effective linking of data from different sources and resumes the idea that the benefits of data linking data can be leveraged in other ways if data are open.
Disregarding (one of) these principles jeopardizes the opportunity to make data interconnected and, hence, “limits the ways it can later be reused in unexpected ways. It is the unexpected re-use of information which is the value added by the web” (Berners-Lee, 2006).
From this, we can see that principles, in general, are fundamental requirements that when not observed can compromise the desired objectives. Movements arise, then, to promote and encourage actions that reflect the principles. In this case, as Ninin (2018) stated, the Linked Data principles, in consonance with the major goals of the Semantic Web proposal and with the Open Data principles, culminated in the Linked Open Data.
The Linked Open Data movement, then, is about applying Linked Data principles to open data (Berners-Lee, 2006) and aims to publish structured data on the Web. In this way, the published data is not only available for viewing and access, but it is structured in a way that allows its effective manipulation, use, and reuse.
As a way of encouraging and evaluating the quality of data published according to the Linked Open Data movement, Berners-Lee (2006) also developed a five-star ranking for publishing and linking data, the 5-Star Linked Open Data. He defends that it is good to publish data openly, using open formats and public standards and that it is better if it links to other data as well.
★ make your stuff available on the Web (whatever format) under an open license
★★ make it available as structured data (e.g., Excel instead of image scan of a table)
★★★ make it available in a non-proprietary open format (e.g., CSV instead of Excel)
★★★★ use URIs to denote things, so that people can point at your stuff
★★★★★ link your data to other data to provide context
LOD and FAIR
As LOD is inherently interoperable, it has the potential to play a fundamental role in the implementation of the I in FAIR.
The number of databases that use the LOD protocol may still be modest compared to other technologies (Neurohackweek & ReproNim, 2016). However, regardless of using LOD or not, it is crucial to “ensure that databases are designed specifically for the web and for reuse by humans and machines” (Neurohackweek & ReproNim, 2016). The FAIR principles offer some guidance to manage data in spite of the type of technology used. Nevertheless, we should emphasize that the FAIR principles are not a protocol or a standard, different from Linked Data. The Linked Data, however, protocol can be fully FAIR, if implemented properly
This 5-star scheme by Berners-Lee (2006) gives fine-granular control over data, optimizes its access, and makes data more discoverable. The table below, based on Hasnain & Rebholz-Schuhman (2018), shows what users are able to do at each level. The 5-stars scheme is incremental and it helps to support data openness by means of better structure, better interoperability, and better reuse.
Table 1. What users are able to do when the 5-stars scheme is employed
|Stars||The user is allowed to|
|★||access the data, consume the data, store it locally, manipulate the data, share the data|
|★★||process the data, aggregate data, perform calculations, visualize the data, export the data into another (structured) format|
|★★★||manipulate the data without any proprietary need|
|★★★★||data can be linked using URIs, data can be partly accessed|
|★★★★★||More data can be discovered while consuming data, the data scheme can be learned directly, users have to deal with broken data links, users have to be cautious when linking new data with existing data because they could have trust or provenance related issues|
Hasnain & Rebholz-Schuhman (2018) compared the 5-star scheme for LOD and the FAIR principles and concluded that the LOD scheme provides a guideline for data providers and publishers to make data more accessible. The FAIR principles, on the other hand, can be applied to non-data assets as well (e.g. codes, workflows, etc.). Figure 1 compares both sets of principles.
Figure 1. Comparison between LOD 5-star scheme and FAIR Principles.
LOD principles’ main objective is data interoperability. FAIR principles aim at reusability. Whereas LOD mandates open data, FAIR requires a stated license for access. FAIR emphasises the need for metadata to improve the reuse of the data. For LOD principles, such metadata would also be considered interoperable data. A key element of LOD principles is URIs. FAIR, on the contrary, allows for a broader range of identifiers.
Finally, FAIR requires a supported IT infrastructure for the achievement of FAIR data, which is understandable since data is not necessarily open and it cannot be reused without a license agreement. Neither LOD nor the FAIR principles do not suggest any specific standard, technology, or solution. Both constitute a high-level guide for data producers and publishers.
Berners-Lee, Tim. 2006. “Linked Data – Design Issues.” July 27, 2006. https://www.w3.org/DesignIssues/LinkedData.html.
Hasnain, Ali, and Dietrich Rebholz-Schuhmann. 2018. “Assessing FAIR Data Principles Against the 5-Star Open Data Principles.” In The Semantic Web: ESWC 2018 Satellite Events, edited by Aldo Gangemi, Anna Lisa Gentile, Andrea Giovanni Nuzzolese, Sebastian Rudolph, Maria Maleshkova, Heiko Paulheim, Jeff Z Pan, and Mehwish Alam, 11155:469–77. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-98192-5_60.
Neurohackweek & ReproNim. 2016. “Introduction to the Web of data”. http://www.repronim.org/module-FAIR-data/01-Web-of-Data/.
Ninin, Débora Marroco. 2018. “Linked Open Data em coleções de patrimônio cultural: aspectos da representação da informação para Humanidades Digitais.” Universidade Federal de São Carlos. https://repositorio.ufscar.br/handle/ufscar/10538.
Feature image: Creative Commons Attribution License