Ratan "Are we there yet? Keeping the promise of open science"
0 likes242 views
This document summarizes the key points from a presentation on open science. It discusses how COVID-19 accelerated open science practices. Major policies now require publicly accessible research outputs and data from federal funders. Implementing open science requires investment in data curation and standards to ensure interoperability and reuse. Case studies show engaging stakeholders and assessing current practices are important initial steps for institutions. Commercialization of research infrastructure and data poses risks if not addressed. Standards and best practices are needed to realize open science's potential and avoid chaos.
Ratan "Are we there yet? Keeping the promise of open science"
- 1. Are we there yet? Keeping the promise of open science Kristen Ratan NISO Technology Summit - 2023 Principal, Strategies for Open Science (Stratos) CoFounder, Incentivizing Collaborative Open Research (ICOR)
- 2. Implementing Open Science Research on Open Science
- 3. Evolving Policy Landscape Tracking the moving targets of Global, US Federal, Funders, and Institutions
- 4. •Free and equitable access •Accelerates discovery and outcomes •Increases reproducibility •Enables reuse •Fuels collaboration •Multiplies impact •Reduces costs •Ensures persistence and legacy Why open scholarship - hopes and dreams
- 5. Created by Matt Lewis, ASAP
- 6. Covid-19 has been a forcing function for open science • Publishers opened access • Researchers shared data • Preprints were rapidly posted • Funders shifted focus
- 8. The Nelson Memo (OSTP ) • US federal funders require free, public access to taxpayer funded research… ○ Without embargo ○ Machine-readable ○ Adequate metadata ○ Includes underlying data Why now? August 2022
- 9. Data sharing - not as easy as it sounds The I and R of FAIR will take significant investment
- 10. Challenges 1. Data curation to ensure adequate metadata 2. Finding appropriate repositories with PIDs 3. Moving from posting datasets to FAIR sharing - interoperability 4. Fueling data lakes and hubs instead of data graveyards 5. Tying code to data for containerized reuse 6. Redistributing credit to track and rewards all outputs and activities
- 11. Graveyards to Lakes to Hubs Graveyard: Isolated datasets in random warehouses and repositories Lake: Disparate data types and formats co-localized or even merged Hub/Commons: Harmonized data with tools, upload, workspaces, and collaboration environments
- 13. Neurodata without borders: open format
- 14. Case Study: Changing Impact and Incentives Aligning Science Across Parkinson’s (ASAP): an open science funding initiative
- 16. Mandatory Preprint Open Access Output sharing Attribution Living DMP
- 20. Getting Started: A University Case Study A UCSC program launching Fall 2023
- 21. Fall 2023: UCSC Celebrates the Year of Open Science Open Scholarship… ○ Is good for humanity and the planet ○ Broadens the reach of your work ○ Is supported by new U.S. government funding mandates ■ Requiring research data deposits ○ Is for all disciplines ○ Improves bibliodiversity and equitable access Join the Library, IT & Research offices in this campaign
- 22. Open Scholarship at UCSC Open Scholarship Strategic Plan • Policy clarification & recommendations • Baseline analysis of open science among UCSC grantees • Campus-wide survey Implementing a Thriving UCSC Open Scholarship Ecosystem • Communication & education • Guidelines & best practices • Preferred repositories & other partners engaged • Support & help for grantees Measuring Ongoing Impact • Establish UCSC compliance workflow & tool chain • Compliance & impact data for annual analysis • Measure against success factors
- 23. Establishing a baseline To assess the current state of data sharing we used DataSeer.ai which machine-reads articles and reports on: 1. How many generated data 2. Among those, how many shared datasets online 3. The repositories they are in, sorted by most commonly used 4. The types of data generated 5. The subject areas / disciplines / departments 6. Interoperability of the data formats generated and shared* 7. FAIRness of repositories* *Coming in 2024
- 24. Assessing the status quo at UCSC In this preliminary sample of ~300 articles, 60% had data shared in a third party site. Data Source: DataSeer.AI
- 25. Code sharing at UCSC DataSeer.AI also looks for code generated and found that, for those articles generating code, 44% had data shared in a third party site Data Source: DataSeer.AI
- 26. Back to UCSC: Repositories used The top repositories show higher Github use, indicating concentrations of code-sharing Data Source: DataSeer.AI
- 27. Next steps • Deeper analysis of data sharing within different departments and disciplines • Survey on attitudes towards and knowledge of open scholarship • Interviewing early adopters on campus and creating case studies • Launching an education and awareness campaign • Designing the implementation plan with preferred tools and services • Determining metrics for success and how they will be measured
- 28. The Cautionary Notes When we don’t control our tool chain
- 29. Where there is innovation… • Commercialization of content Commercialization of data • Proprietary platforms and paywalls are strong • Open source alternatives are fractured and under-resourced • The new policies will lead to a tidal wave of sharing • The trusted and interoperable tool chain isn’t in place
- 30. Most of academic infrastructure is closed The 2021 SPARC report identifies three main concerns: 1. Insertion of tracking software in services sold to academic libraries 2. Collection and sale of data by some commercial vendors with ties to the academic community to governments and law enforcement 3. Risks and inequities of online exam proctoring tools https://infrastructure.sparcopen.org/media/posts/report-landscape-analysis-2021-update.pdf
- 31. Commercial Platforms as data brokers Large commercial platform and publishing companies have been buying up the technologies that house the content, people, and workflow data of the scholarly process from end-to-end. They are consolidating data on: - What researchers are working on - What students are studying - Trends in research and collaboration pre- and post- publication - Products, equipment, and services in development and use - Spread, reach, and influence of research on other researchers, the public, the markets, and policy
- 32. AI-conducted research As AI increasingly conducts analyses, the closed nature of it means we’ve lost reproducibility that we hoped to gain from open policies
- 33. Is open scholarship increasing equity? We hear this claim We have little to no evidence This is an area we need to study
- 34. Standards are the root of all good We need standards and best practices to avoid chaos
- 35. • Common workflows and pipelines • Standards such as: • Parameters for persistent identifiers • Minimum viable metadata • Methodologies for tracking openness • Compliance measures • New metrics for excellence Standards and Best Practices