Interaction between OBIS/Unesco with TDWG and DISSCO on Schema.org mapping to TDWG standards

[pmergen]

Based on a discussing of addressing thematic siloing, would need to interact with TDWG and their biodiversity standards, Schema.org as well as with DiSSCo to open in practice to other platforms and other disciplines. Plans would to test with OBIS/UNESCO . Link it also to the reviving of the RDA Biodiverity group to add OBIS to it. involved: Pier Luigi Buttigieg Sam Leeflang Sharif Islam Patricia Mergen

back ground reading : https://worldfair-project.eu/2023/12/05/worldfair-new-interoperability-specifications-and-policy-recommendations-d11-2/

[jmckenna]

thanks @pmergen. Did you mean above "Plans would be to test with ODIS/UNESCO" ? (I wasn't sure if you swapped ODIS with OBIS in that sentence)

[pmergen]

in the discussion ODIS and OBIS was mentionned ..

[pbuttigieg]

@jmckenna @pmergen the aim here is for information standards organisations like TDWG to offer authoritative maps between the standards they operate (e.g. TDWG's DwC and schema.org)

We set an example of a machine-actionable map using SSSOM between DwC and the Genomic Standards Consortium's MIxS standards here https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10565567/

If TDWG can maintain such a map to schema.org, DCAT, and other generic standards (naturally, only covering the generic content), then systems like ODIS could steam ahead with biodiversity data harmonisation when it interacts with systems like DiSSCo

There's an RDA group exploring semantic maps (@fils mentioned in our CDIF meetings ), and they may be interested too. However, this shouldn't drag out - we know SSSOM mappings work well enough now.

The challenge will be more in the syntactic variation in value space than they key/property space.

[sharifX]

@pmergen @pbuttigieg

Thanks for initiating this discussion. From DiSSCo's and the overall biodiversity data integration and FAIR points of view, we are very much interested in this. We are also involved in some of the mapping and SSSOM discussions, so that alignment is beneficial (see a recent hackathon around SSSOM in the Biodiversity Genomics Europe project). There is also ongoing discussion in the Biodiversity Digital Twin project around this topic.

To give a quick summary, DiSSCo is currently working on an MVP, and we are actively developing our Digital Specimen data model (based on FAIR Digital Object specifications). You can see our schema repository here and on GitHub.

Here's a simple high level mapping example that could be easily done. We can enhance this with SSSOM structure.

openDS (ods) -> scehma.org

• ods:id -> identifier (https://schema.org/identifier)
• ods:version -> version (https://schema.org/version)
• ods:created -> dateCreated (https://schema.org/dateCreated)
• ods:specimenName -> name (https://schema.org/name) Note: this needs to allow both biological and geological specimen names.
• dcterms:license -> license (https://schema.org/license)
• dcterms:modified -> dateModified (https://schema.org/dateModified)
• dcterms:accessRights -> hasDigitalDocumentPermission (https://schema.org/hasDigitalDocumentPermission or https://schema.org/conditionsOfAccess)
• dcterms:rightsHolder -> accountablePerson (https://schema.org/accountablePerson)
• dwc:institutionId, dwc:institutionName -> provider (https://schema.org/provider)
• dwc:recordedBy -> creator (https://schema.org/creator)

However:

Properties such as ods:midsLevel, ods:physicalSpecimenId, ods:physicalSpecimenIdType, dwc:preparations, dwc:disposition, dwc:collectionCode, dwc:collectionId, dwc:informationWithheld, and dwc:verbatimLabel do not have direct equivalents in schema.org.

We are currently working on a more formalised and structured description of these terms to make them more FAIR. As some of these terms are related to Digital Specimen, specimen digitisation pipeline, and museum workflow, and are specific to DiSSCo, we will probably need to maintain the namespace for them. However, we can align with TDWG around the mapping schema.

And yes, syntactic variation in the value space is always challenging.

If this can be achieved, that would be amazing:

"If TDWG can maintain such a map to schema.org, DCAT, and other generic standards (naturally, only covering the generic content), then systems like ODIS could steam ahead with biodiversity data harmonization when it interacts with systems like DiSSCo."

We also need to reach out to various other projects discussing mapping, like FAIRCORE4EOSC's metadata schema cross walk registry.

[pmergen]

Input from Ben Norton :

Two short notes to append.

1. Latimer Core is already mapped to schema.org.
2. Schema.org is a 'high-level' schema for generally applicable concepts. Several groups within the schema.org space are actively extending the schema to more targeted areas, such as Taxon (https://schema.org/Taxon), but it remains more generalized than most things under the TDWG umbrella. Since most concepts will be narrower matches to terms in schema.org, I think it's questionable if there are significant interoperability advantages to spending significant resources on this topic. Regardless, mapping to schema.org is a relatively straightforward process. I don't think it requires the establishment of a group; rather, someone just needs to do it.

[wouteraddink]

@pmergen can you add me too?

[pmergen]

Yes. It is an open GitHub . I think you cab access it and answer to the tread.

I tried to add Ben but his Github name would not appear in the list..

Pat

On Mon, 3 Jun 2024, 12:50 Wouter Addink, @.***> wrote:

@pmergen https://github.com/pmergen can you add me too?

— Reply to this email directly, view it on GitHub https://github.com/iodepo/odis-arch/issues/434#issuecomment-2144885003, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC4XITB3CI6LZ6HEQ2UT77DZFRC7BAVCNFSM6AAAAABIS6CPV2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBUHA4DKMBQGM . You are receiving this because you were mentioned.Message ID: @.***>