Project kick-off
2025-03-10
Rick Gilmore
Donald Rumsfeld’s unknown-unknowns and the Johari window, Wikipedia contributors (2025)
Agenda
- A brief history
- Overview of Aims
- Next steps
A brief history
Databrary.org is a digital, non-commercial, academic data library housed at New York University (NYU). It is the world’s only repository specialized for storing, streaming, and sharing research video and linked identifiable data at scale.
- NSF workshop (2011)
- NSF (2012-2018); NICHD proposal (2013-201)
- DB1 go-live (March 2014)
- Alfred P. Sloan Foundation (2017), James S. McDonnell Foundation (2018) grants
- NIH Play & Learning Across a Year (PLAY) (2018-)
- John Templeton Foundation
- DARPA
- NSF OAC-2032713 (2020-2023)
- NYU internal award (2023, 2024)
- Gates Foundation National Tutoring Observatory (NTO) subaward (2025-)
- NSF BCE-2444730 (2025-2027)
- DB2 go-live (March 2025)
DB1 growth through Feb 2025
Volume sharing status through Feb 2025
Citations through Feb 2025
Values
- As open as possible
- As closed as necessary
Vision
- Video is an unparalleled source of data, documentation, and demonstration
- Make Databrary the leader in open behavioral science
- A tool for discovery
Overview of Aims
Aim 1: Enhanced data discovery
- Why: Accelerate data reuse
- How:
- Enhanced search
- Annotation layers: Search & visualize
Aim 2: Custom collections
- Why:
- Make data reuse easier
- PLAY Project release
- How:
- Select, clone/copy shared files, sessions, metadata, volumes
- Maintain virtual links
- AKA “virtual volumes”
Aim 3: Workspaces
- Why:
- Active curation (Soska et al. 2021) \(>>\) post hoc
- In progress \(\neq\) published/shared
- How:
- Workspace with essential DB2 metadata
- Flexible views
- Semi-automated curation for sharing
Aim 4: Scriptable access
- Why: Transparency & reproducibility
- How:
- DB2 API
- Secure authentication
- Update databraryr
- Develop & publish databrarypy
Aim 5: Curation
- Why: Self-curation idiosyncratic
- How:
- Review fully shared, overview only, private
- Curate
Timeline
Other opportunities
- Overhaul, polish UI/UX
- Institutional subscriptions (NYU TOV TAC 2025)
- Institution dashboard, analytics
- databrary.ai/videobrary.ai
- {zoo,clinic,teach}brary
- Web-based video annotation
Discussion
- What do we know?
- What don’t we know?
- What’s important?
- What’s urgent?
Known-knowns
| Urgent | Not Urgent | |
|---|---|---|
| Important | Upload; search; Montrose | custom collections1; databrary{r,py}; curation |
| Not important | UI cleanup |
Also important/not urgent
- Institutional subscriptions
- Workspaces \(\rightarrow\) shared volumes
- See RIDIR 2018 Aim 2
- Spreadsheets: Use cases, interfaces, prior work
- Indexing, searching, & visualizing annotation files
- See RIDIR 2016 Aims 1, 2
- Work with TalkBank.org on CHAT integration?
Resources
Code
These slides were written in R Markdown and rendered using Quarto to HTML slides using the reveal.js framework.
References
Soska, Kasey C, Melody Xu, Sandy L Gonzalez, Orit Herzberg, Catherine S Tamis-LeMonda, Rick O Gilmore, and Karen E Adolph. 2021. “(Hyper)active Data Curation: A Video Case Study from Behavioral Science.” Journal of Escience Librarianship 10 (August). https://doi.org/10.7191/jeslib.2021.1208.
Wikipedia contributors. 2025. “There Are Unknown Unknowns.” Wikimedia Foundation, Inc. February 11, 2025. https://en.wikipedia.org/wiki/There_are_unknown_unknowns.
