‘Making your research data FAIR accelerates scientific discovery’
FAIR - Findable, Accessible, Interoperable, Reusable - data enhance machine readability.
10/08/2019 | 11:53 AM
As a postdoctoral researcher at the Knowledge Representation & Reasoning Group of the Vrije Universiteit Amsterdam Albert Meroño Peñuela is an enthusiastic promotor of FAIR data. ‘The amount of data available on the internet has grown immensely over the past few years. However to find, combine, access and reuse them is very difficult. We lose a lot of time ‘wrangling’ the data: making them machine readable and ready to analyse. Therefore it is important that researchers make their data Findable, Accessible, Interoperable and Reusable.’ What does that mean in practice?
FAIR data benefit society
‘In Artificial Intelligence research 60% of the time is now spent on preparing the data for the computer and in the Digital Humanities it is even worse: it takes up to 80% of the time’, says Meroño. ‘Artificial Intelligence can be applied in many different domains: from Medicine, Pharmacology and Bio-informatics to the Digital Humanities. Health data for example can help to: detect and diagnose disease more effectively and speed up drug discovery. Meroño’s own domain of application is Digital Humanities and in particular History & Musicology. ‘Making your data FAIR helps the machine to find and analyse them and thus accelerates scientific discovery. It benefits society.’
FAIR datasets are becoming more important in the assessment of research(ers)
According to Meroño there are also personal benefits to making your data FAIR. ‘It contributes to your academic reputation and impact. At the moment there is a change in the way research and researchers are assessed . How much others are using your FAIR-datasets is becoming more important in the evaluation of research. If you make your dataset FAIR it becomes easier to measure your factual impact by monitoring the use of the data.’ FAIR data can also contribute to attracting new research partners. Meroño has recently started to publish his FAIR datasets as academic publications. “In those papers I document my datasets and those get cited. So making my data FAIR has already brought me some benefits.”
‘In our Knowledge Representation group we think about the best way to make data understandable for the machine. I investigate the integration of and access to diverse types of knowledge, with a special interest in Digital Humanities data and workflows. I am interested in knowledge graphs, Linked Data, Web query languages, and APIs.’
Five recommendations for FAIR data
Meroño has five recommendations for researchers interested in making their data FAIR.
1. ‘Have a look at the Linked Data website.’ It offers a list of technical recommendations and best practices for exposing, sharing, and connecting pieces of data, information and knowledge on the Semantic Web using URIs and RDF. If you apply those you are working FAIR.’
2. Carefully think of the identifiers, such as URLs or URIs, you use for your dataset. URLs locate web documents; URIs can locate anything. One identifier for your dataset is not enough! It is important to also give identifiers to columns, persons and places. Think carefully of the entities in the real world that are represented in your data.
3. Use HTTP (hypertext transfer protocol) to make your data accessible. We are now often confronted with a login screen or a paywall when we try to access data, and then the computer can not read the data. We should of course not forget about authentication and privacy but there are ways to organize authentication in compliance with HTTP.
5. Use a clear data usage license.
Discipline specific implementation of FAIR
The FAIR principles are not discipline specific. However, the implementation of FAIR needs to be discipline specific. With the help of researchers Library Carpentry makes brief guides with short descriptions related to FAIR within several disciplines: librarycarpentry.org/Top-10-FAIR . Meroño took part in a workshop for formulating the Top 10 FAIR Things for Music . The Library of the VU Amsterdam was involved in formulating the Top 10 FAIR Things for Astronomy . The Community Manager RDM, Maria Cruz, participated in a workshop in May that involved researchers and librarians from research institutions around the world.
These are the disciplines for which Library Carpentry with the help of researchers has made Top 10 FAIR things - lists.
- The European Open Science Cloud
- Research Software
- Research Libraries
- Research Data Management Support
- International Relations
- Humanities: Historical Research
- Biomedical Data Producers Stewards, and Funders
- Australian Government Data/Collections
Want to know more? Have a look at: