1 of 25

Strategies of tackling barriers to research data publishing:

A case study from the University of Michigan Library

Jake Carlson

Director of Deep Blue Repository & Research Data Services

jakecar@umich.edu

2 of 25

Background

The University of Michigan Library supports two institutional repositories:

Deep Blue Documents (launched in 2006)
Deep Blue Data (launched in 2016)

3 of 25

Our Mission

Provide a means for a means for the U-M community to make their work openly accessible to anyone in the world.

Publish content of scholarly or educational value, research data in particular, that may not have a defined path for dissemination.

4 of 25

Data Curation

Metadata Entry

Upload Data Files

Terms of Use

Submit

Check files

Understand documentation

Request missing information

Augment the submission

Transform the format

Evaluate for FAIRness

Document throughout

5 of 25

Data Management Across the Lifecycle

6 of 25

Prepared to Publish?

Data are in different states of readiness when they are deposited to Deep Blue Data.

7 of 25

8 of 25

Ethical. Reusable. Better.

datacurationnetwork.org

Strategy for Data Publishing #1

Data Management Plans as an Outreach Tool

Ethical. Reusable. Better.

datacurationnetwork.org

9 of 25

Accessing U-M’s DMPs

Working with U-M’s Office of Research, I was able to negotiate courtesy access to our grants management system.

This gives me the means to view DMPs from awarded grants.

10 of 25

Connecting with future depositors

Reviewing DMPs allows me to identify who intends to deposit data with us and reach out to them to introduce them to our services.

Photo by ThisisEngineering RAEng on Unsplash

11 of 25

Metadata and Documentation

	NIH n=233	NSF n=173	DoD n=11	NASA n=22	Other n=22	Total N=461
Metadata - Mentioned	14 (6%)	23 (13%)	2 (18%)	4 (18%)	2 (9%)	45 (10%)
Metadata - Detailed	3 (1%)	41 (24%)	2 (18%)	3 (14%)	4 (18%)	53 (11%)

Documentation - Mentioned	33 (14%)	43 (25%)	2 (18%)	5 (23%)	3 (14%)	86 (19%)
Documentation - Detailed	12 (5%)	40 (23%)	2 (18%)	1 (5%)	7 (32%)	62 (13%)

12 of 25

Ethical. Reusable. Better.

datacurationnetwork.org

Strategy for Data Publishing #2

The Data Curation Network

Ethical. Reusable. Better.

datacurationnetwork.org

13 of 25

Mission

Trusted, community-led network of curators advancing open research by making data

Ethical. Reusable. Better.

datacurationnetwork.org

14 of 25

Ethical. Reusable. Better.

datacurationnetwork.org

Curate data as a cross-institutional network of nearly 50 individual experts �

Offer professional development opportunities for an emerging data curator professional community�

Create and openly share data curation best practices�

Informal research teams addressing a specific topic: big data, human subjects, racial justice...

�Build community at annual events for discussion, training, and networking.

Curation

Education

Primers

Interest Groups

Community

Ethical. Reusable. Better.

datacurationnetwork.org

Ethical. Reusable. Better.

datacurationnetwork.org

15 of 25

Ethical. Reusable. Better.

datacurationnetwork.org

Ethical. Reusable. Better.

datacurationnetwork.org

Ethical. Reusable. Better.

datacurationnetwork.org

In practice, here is what the shared staffing curation workflow looks like– on the left hand side we have the DCN Partner institution who receives data from a researcher– the local curation team triages the dataset to decide whether they would like to curate it locally, or if they would like additional support from a DCN expert. When it is sent to the DCN, the coordinator matches the dataset with the pool of available curators, depending on discipline and format considerations. A DCN curator then applies the CURATE(D) workflow, that I described earlier and provides feedback and recommendations to the local institution, who closes the loop on their ends.

Researchers deposit like normal – repository staff appraise the content, ingest it as appropriate
DCN functions as a microservice layer (the “human layer in your repository stack”)
Local institution maintain full responsibility for all technical functionality (eg. storage) and authority for local decision-making (what to ingest, how long to retain, etc.)
Seamlessly integrates into all repository systems (Samvera, Fedora, DSpace, etc.)

16 of 25

Data Curation Primers

The project began as a capstone to our Specialized Data Curation Workshops that were generously funded by the Institute of Museum and Library Services (IMLS).
27 primers released so far on Github!

Grant # RE-85-18-004018

Peer-reviewed concise resources to assist the data curator with recommendations for curating specific formats or curation topics!

Free and

open!

Ivey, Susan; Koshoffer, Amy; Sneff, Gretchen; Wang, Huajin. (2019). Confocal Microscopy Images Data Curation Primer. Data Curation Network GitHub Repository.

Ethical. Reusable. Better.

datacurationnetwork.org

17 of 25

Ethical. Reusable. Better.

datacurationnetwork.org

Strategy for Data Publishing #3

Cultivating Allies at U-M

Ethical. Reusable. Better.

datacurationnetwork.org

18 of 25

Public Access to Research Data Working Group @ U-M

Charged by the Provost and V-P for Research in 2019 to craft recommendations for supporting U-M researchers in making their data publicly accessible.

Membership included the Library, Research IT, Statistical Consulting, the Office of Research, the Med School, ICPSR, and General Counsel’s Office.

19 of 25

Framing the Issue

As a part of the mission of the University
As a potential strategic advantage

20 of 25

Three Interrelated Challenges

A lack of incentives for researchers
The scale and heterogeneity of research data at U-M
The lack of common understandings and connections between units providing research support

Source: Schneider, Sarah L. 2018 FDP Faculty Workload Supplement: the University of Michigan. October 2019.

21 of 25

Recommendation One

U-M should create a faculty committee knowledgeable on the challenges of sharing research data and charge them with acting on the issues identified in this report.

Initial areas of responsibility:

Develop an understanding of researchers experiences across U-M in responding to data sharing requirements
Develop a promote a shared set of institutional values and norms at U-M around data sharing
Address gaps and inadequacies of U-M’s policies with research data

22 of 25

Recommendation Two

The units that provide support for data management, sharing, curation, and preservation at U-M should coordinate and align their work even more closely.

Possible areas of responsibility:

Inventory of services
Better inventorying of U-M’s research data
Developing machine actionable Data Management Plans
Considering storage needs for data sharing and preservation
Offering educational programming

23 of 25

https://research.umich.edu/research-data-stewardship/

24 of 25

Concluding Thoughts

Data publishing at scale requires:

A means to connect and communicate with data authors
A robust infrastructure and peer community to support the work
Broader institutional awareness and support

25 of 25

Thank you!

Jake Carlson

Director of Deep Blue Repositories and Research Data Services

jakecar@umich.edu