Webinar: FAIR Vocabularies in Population Research, 12 June, 12:00-13:30 UTC

Monday 12 June, 12:00-13:30 UTC

Register: https://us02web.zoom.us/meeting/register/tZYvcu6hpzouGtTdnZLq2zMtDSuw4d24cihS

This event will launch the report ‘FAIR Vocabularies in Population Research’ https://doi.org/10.5281/zenodo.7818157 produced by the joint IUSSP-CODATA WG with the same name.  The speakers will present the report and its implications and discuss the next steps for implementing its recommendations.

IUSSP and CODATA co-sponsored a working group to study how population research can benefit from the rapidly developing standards and technologies associated with the FAIR principles that all data should be “Findable, Accessible, Interoperable, and Reusable” by both humans and machines (Wilkinson et al., 2016).

Demography is an empirically focused field with a long tradition of widely shared, easily accessible, data collections.  FAIR vocabularies, which allow machines to associate data with concepts, can save researchers hours of tedious work by automating processes of data discovery and harmonization.  The report introduces readers to international standards for documenting data (metadata) that underlie international infrastructures for producing and disseminating demographic data, and it recommends enhancing these services through application of the FAIR principles.  The report builds on the “Ten Simple Rules for making a vocabulary FAIR”  (Cox et al., 2021), prepared by a group formed at a workshop convened by CODATA and DDI to describe how a FAIR vocabulary will work with international standards for documenting and sharing social science data.

The working group calls for IUSSP to create a FAIR Vocabulary of Demography.  Online vocabularies including demographic terms already exist, and most of them define key terms in ways incompatible with demography.  Population research will be at a disadvantage without an authoritative FAIR vocabulary of its own.  Fortunately, a new FAIR Vocabulary of Demography can build upon IUSSP’s long history of support for dictionaries of demography in multiple languages.

Speakers:

• George Alter, University of Michigan
• Abdulla Gozalov, United Nations Statistics Division
• Steven McEachern, Australian Data Archive, Autralian National University

Section 10.3 of the report provides specific recommendations for the DDI Alliance, including:

1. We encourage the DDI Alliance to develop guidelines and practices for registries of vocabularies, variables, and value domains similar to the development occurring in the SDMX world. DDI registries should work to encourage reuse of survey instruments and coding schemes in ways that will make it easier to discover and harmonize data. Since the DDI community is very decentralized, we expect that DDI registries will be federated using the DDI Agency Registry to simplify the assignment of persistent identifiers.As an XML based standard since its inception, DDI has supported URNs but as new serializations emerge, support should be accelerated for RDF URIs that are natively resolvable to locations on the Web.
2. IUSSP strongly recommends that data producers apply URIs when data is created. It is much more efficient for persistent identifiers to be assigned by data producers, and these identifiers will provide more accurate information than annotations applied by data repositories late in the data life cycle. Tools for assigning persistent identifiers throughout the data lifecycle are already available. Wherever possible, these URIs should point to existing FAIR vocabularies, such as ELSST and the future FAIR Vocabulary of Demography.
3. The DDI Alliance should also take other steps to facilitate harmonization of data sets. The assignment of persistent identifiers described in DDI registries is a pre-condition for reducing the costs of data harmonization, but other steps, such as using tools like SKOS to map variables to concepts are needed. We strongly encourage the DDI Alliance to contribute to research and the development of tools for automating linking identifiers across registries discussed below.