talk · Thursday 24 September · 09:55–10:10 · SAL A
A framework for aligning collections data with the global data ecosystem from data creation to extension in the Smithsonian National Museum of Natural History Department of Paleobiology
Holly Little · From Cabinets to Clouds - Bridging the gap between collection- and database management
Summaries and key moments are machine-produced. Check the recording for the speaker’s exact words.
The short versionMaking paleo collections work in the global data ecosystem needs an iterative, full life-cycle framework, not one-off digitisation.
↶Go back to 2025Data Package pilot · 1 connection↓
Revisit Living Data 2025 alongside this talk. These connections were checked against captions; related examples may come from different teams rather than a direct follow-up.
Characterizing paleontological specimens for maximal alignment with global data practices
While that is true, it will take work to implement. As I noted in our caveats, we had to map our data manually. Most collections management systems do not have the fields or tools to store or generate some of the necessary data that makes the DARO data package a powerful publishing model. In particular, the extensive and important use of identifiers is something our community doesn't quite have the ability to meet yet. And as we all see with within our communities, the resources and expertise to adapt to a new system can vary widely across collections and institutions. To help meet the needs of the Paleo collections community, we're working with the Paleo data working group to develop a library of shared resources and tools in a new web resource we're calling the Paleo data knowledge hub. We are also working to create a Darwin core implementation guide that is specific to fossils from Paleo collections and will address the Darwin data package once finalized. We're also working to create example records for the various fossil types and data groups which can be used to illustrate the full data mobilization pipeline alongside the implementation guide. And with that, thank you. We
2026 This talk
In the later paleo framework Holly recalls a mapping exercise “last year” and argues for different DwC-DP publishing profiles for different kinds of material.
A framework for aligning collections data with the global data ecosystem from data creation to extension in the Smithsonian National Museum of Natural History Department of Paleobiology
does a lot to give us a more dynamic way of thinking about how we map and model the data coming out of paleo collections. And because we do have such a variety of material, it turns out that it isn't actually as useful to use one model for every single thing in a paleo collection. And last year we did a mapping exercise as they were testing Darwin Core data package, and found that we could leverage different modeling profiles or publishing profiles to better represent different types of material, whether it is something that really relies on relationships between organisms, or we have cores that cross many different geological horizons. So better leveraging the more structured data models that we now have to map our data and get the right data points out for researchers for discovery. This also includes leveraging the Darwin Core material entity type to be more specific about what our records represent, and incorporating all the work we did in enhancement to include links and persistent identifiers and building out our Darwin
The later talk develops a practical approach to heterogeneous fossil data rather than requiring one mapping for every specimen type.
About this connection
The reference to last year supports continuity of the workstream, but does not identify this exact 2025 presentation. It does not establish that earlier software or identifier barriers have been solved.
This pilot covers one 2025 symposium and selected 2026 moments. Automatic captions can contain mistakes; audio has not been checked.
Overview
What this was about
Holly Little presented the Paleobiology department's new five-year, gift-funded 'paleo digital framework' for the world's largest fossil collection. The collection holds about 42 million objects, 800,000 records, and needs an estimated 10–13 million records. The framework covers the data life cycle from capture to enhancement, extension and publishing. It uses imaging-first capture, adaptable specimen surveys for core data points, material entity type and MIDS, and verbatim preservation before AI text extraction. It also aligns taxonomy and geological context with external sources and feeds corrections back to them, and uses Darwin Core Data Package profiles. The informatics team grows from one person to five. She asked the community for workflows and challenges to test.
Why it matters. It is a concrete institutional roadmap for implementing years of community standards work (material entity type, DwC-DP, geological context). The team explicitly offers to pilot and share solutions.
Key ideas
In the room
A 2018 analysis showed low use of Darwin Core terms in paleo collections due to misunderstandings, missing terms and unlinkable resources.
The NMNH Paleobiology collection has about 42 million objects and about 800,000 records; an estimated 10–13 million records are needed.
The informatics team was one person until 2025, then two, with three more staff funded by a gift for a five-year project around 1,000 significant objects.
Principles: work iteratively, image first, share data widely and often, don't wait for perfection, align with global standards and FAIR.
Capture only core data points per specimen type and reuse ledgers (e.g. a locality shared by thousands of specimens) instead of re-transcribing labels.
Enhancement aligns taxonomy (WoRMS, Paleobiology Database, Catalogue of Life) and geological context with external sources and feeds issues back.
DwC-DP mapping showed that different publishing profiles suit different material types (e.g. organism relationships vs cores crossing horizons).
Jump in
Notable moments
In their words
Transcript
Automatically generated captions can contain mistakes, especially in names and technical terms. Times are relative to the room recording.
