Because good research needs good data

DMPRoadmap Team at the maDMP Hackathon

Maria Praetzellis | 12 June 2020

Research Data Alliance (RDA) recently hosted a three day (27-29 May 2020) machine-actionable DMP hackathon to build integrations and test the Common Standard for maDMPs. The event, coordinated through teams at RDA-Austria and TU Wien, was well attended with over 80 participants from Australia, Europe, Africa, and North America. 

The teams that work on DMP Tool (dmptool.org) and DMPonline were really pleased to represent our shared DMPRoadmap codebase and show our conformance with the standard and ability to exchange DMPs across systems. This blog post details the work of the DMPRoadmap group in the hackathon, for a full review of all outputs please visit the Hackathon GitHub.

What did we work on?

Maria Praetzellis and Sarah Jones, product managers from DMPRoadmap, joined the hackathon “TigTag” team and focused on mapping maDMPs to funder templates. During the hackathon, their group successfully mapped required questions from several funder specific DMPs including: 

  • Horizon 2020
  • Science Europe
  • National Science Foundation
  • U.S. Geological Survey

The goal of the exercise was to develop guidance on how to normalize the ways that fields from specific funder templates can be mapped to the standard, and, when necessary, develop extensions to incorporate template specific needs. The team came up with several proposals for changes to the documentation and structure of DMP Common Standard and made a few recommendations for extensions to the standard. The team is now assembling the recommendations and will submit ideas as issues to the Common Standard GitHub so work can be tracked going forward. 

Brian Riley and Sam Rust, developers from DMPRoadmap,  joined the hackathon “DMP Exchange team” and worked to determine how the RDA Common Standard JSON format could be used to exchange DMP metadata between tools. Their team provided a staging service and granted API keys to other development teams to allow testing of prototypes, which helped all participants debug issues. Over the course of the hackathon, our new maDMP API helped developers of the following DMP systems implement their own APIs:

Based on this work, we were able to exchange maDMP metadata between DMPTool and those three systems by the end of the hackathon.  Below are screenshots of DMP exports from the Data Stewardship Wizard that were imported into the DMPTool. Because we were each using the RDA Common Standard format, the new DMP was created within the DMPTool and the appropriate metadata was successfully mapped: title, description, project start/end dates, grant ID, contact information, and contributor information.

While the data models used by many systems do not yet offer full support of the RDA Common Standard model, progress was made towards mapping the high level DMP information across the board. Also, the confirmation that these systems could exchange information using RDA Common Standard JSON was encouraging and will likely open the door for future integrations. 

Other outcomes

We also collaborated with members of the DMP Melbourne, University of Cape Town and Stockholm University on an integration with their institutional repository platform. The teams were interested in pushing both DMP metadata and the physical DMP document into that repository. However, they did not yet support the maDMP standard. So the team created two separate prototype scripts. The first script extracts DMPs from a DMPRoadmap system and creates a placeholder Project that future datasets can be connected to and also uploads a PDF copy of the DMP. The second script converts their JSON into RDA Common Standard compliant JSON. While their institutional repositories do not contain many DMPs at this point, a service like this could help extract DMPs for import into DMP systems that utilize the RDA Common Standards in the future. We hope to build upon this work to facilitate integrations with additional repositories in the future. 

Future work 

Hackathon participants are now collating work produced during the hackathon into a final report. In addition, participants expressed interest in:

  • More communities. Most of the attendees at this hackathon were developers from DMP-focused tools. In the future, it would be great to have participants from other communities, including developers of CRIS systems, data repository platforms, and ethics tools.  This would help us expand the types of use cases being served.
  • More PIDs. The power of connected information replies on persistent identifiers.  We would like to increase our connection with various standards and integrate with the Research Organization Registry (ROR), the Funder Registry, and the Contributor Roles Taxonomy (CRediT) to provide more structured information to support such integrations.

Thank you again to the team at RDA Austria and TU Wein for organizing the hackathon.  If you’re interested in tracking future development and outputs of this work please follow the GitHub and consider joining the RDA Common Standard Working Group or Active DMPs Interest Group. And many thanks to Maria Praetzellis for letting us to re-post the original blog.