Share
Embed
Open Research + Reproducibility
Presentation given at UW Libraries, February 2017.
Brianna Marshall
February 15, 2017
Transcript
53 slides
-
About these workshops Funder Public Access Requirements Wed, Oct. 12,
2:00-3:30 at Steenbock Thurs, Oct. 13, 9:30-11:00 at Memorial Research Data Management & Sharing Wed, Nov. 16, 2:00-3:30 at Steenbock Thurs, Nov. 17, 9:30-11:00 at Memorial Open Access Publishing Wed, Dec. 14, 2:00-3:30 at Steenbock Thurs, Dec. 15, 9:30-11:00 at Memorial Open Research & Reproducibility Wed, Feb. 15, 2:00-3:30 at Steenbock Thurs, Feb 16, 9:30-11:00 at Memorial Authors’ Rights Management Wed, Mar. 15, 2:00-3:30 at Steenbock Thurs, Mar. 16, 9:30-11:00 at Memorial TBD Wed, Apr. 12, 2:00-3:30 at Steenbock Thurs, Apr. 13, 9:30-11:00 at Memorial Library administration effort to support staff work and development around these topics -
Scope • What is open research? • The why and
how of open research • Open research techniques and tools • Openness examples -
what’s up with researchers? Ample technology to generate data but
few skills to manage it effectively Movement toward openness, impacted by OSTP and spurred by early career researcher expectations Disciplinary culture shifts toward data reuse + reproducibility Need for multi-purpose online spaces to collaborate, share, store, and archive research outputs (including data) -
research is increasingly digital. • Multi-institutional • Grant-funded • Shared
infrastructure • Computationally driven Image courtesy of #wocintechchat -
Aspects of Open Research • Project management • Data management
• Workflows • Tracking myriad outputs • Collaboration -
Munafo et al. (2017) A manifesto for reproducible science. Nature.
Why is reproducibility a problem? -
Reproducibility project insights • Original teams shared materials, methods, and
data • Replications preregistered protocols openly • Data, materials, and code shared by replication teams From Erin McKiernan, “Sharing in Science.” (CC BY) -
• Version control • Dynamic documents • Building from source
From Pierce Edmiston: github.com/pedmiston/reproducible-research -
describing openness “For me, openness means an honesty about the
messiness of research and transparency in methods and process that helps both the researcher and the audience. It helps the researcher by allowing others to comment and get involved in the research earlier if they spot flaws in the methodology or process, and the audience by showing (especially junior or first-time researchers) that research is rarely a clean progression from simply-defined goals to a final research output, and instead involves reworking and change as certain aspects of the originally-scoped research may become untenable or new areas prove to be more interesting or researchable.” –Thomas King, Open Research course participant https://courses.p2pu.org/he/courses/2377/content/4682/ -
six laws of open research 1. All data are open
and all ideas are shared. 2. Anyone can take part at any level. 3. There will be no patents. 4. Suggestions are the best form of criticism. 5. Public discussion is much more valuable than private email. 6. An open project is bigger than, and is not owned by, any given lab. opensourcemalaria.org -
how to be open • Share your process – Documentation – Notes,
drafts, etc. – Preregister your project • Share your outputs – Publications – Data -
open dissemination • Create a project identity on various tools
• Make the project blog the core part of our project identity • Share research progress, outputs and methods on a regular basis, via our website/ blogs and other media • Share data openly, including survey results • Share methodology and research instruments under a CC-BY license Adapted from Open Research Course, “Examples of Open Dissemination.” https://courses.p2pu.org/he/courses/2377/content/4695/ -
tools supporting open dissemination • Social networks – Personal – Academic •
Repositories • Collaboration platforms • Portfolios/personal websites -
steps toward reproducible research – Everything with a script –
Organize your data and code – Automate the process – Turn scripts into reproducible reports – Turn repeated code into functions – Package functions for reuse – Use version control – License your software Note: These steps apply primarily to computational research. http://kbroman.org/steps2rr/ -
erin’s openness pledge I pledge to: – edit and review
only for open access journals – publish only in open access journals – openly share my working manuscripts – openly share my code, when possible – openly share my data, when possible – openly share my notebooks, when possible – ask my professional societies to support open research – speak out in support of open research -
- Denne Reed, Homo Naledi team [ http://www.nature.com/news/crowdsourcing-digs-up-an-early-human-species-1.18305 ] “We’re
more interested in openly sharing data… the advantages in collaboration far outweigh any of the risks.” -
“The investigators collected their samples, returned home and published the
startling results in European medical journals. Few Liberians were then trained in laboratory or epidemiological methods. Even today, downloading one of the papers would cost a physician here $45, about half a week’s salary. …To our knowledge, no senior official now serving in Liberia’s Ministry of Health had ever heard of the antibody studies’ findings. Nor had top officials in the international organizations so valiantly supporting the Ebola response in Liberia, including United Nations agencies and foreign medical teams.” - BERNICE DAHN, VERA MUSSAH and CAMERON NUTT http://www.nytimes.com/2015/04/08/opinion/yes-we-were-warned-about-ebola.html -
Open Research + Libraries CU Boulder – Open and Digital
Scholarship Services University of Cambridge Open Research pilot -
What We Can Do Be proactive supporters of open research
beyond enforcing compliance. How? • Immerse ourselves in the tools • Help bridge the technical skills gap • Create targeted programming • Partner with existing platforms -
What We Can Do Right Now! Immerse ourselves in the
tools • Familiarize yourself with the Open Science Framework, Authorea, Jupyter notebooks, and R. Help bridge the technical skills gap • Library involvement and leadership in Software Carpentry and Data Carpentry. Create targeted programming • Presentations and workshops on open research related topics. Partner with existing platforms • OSF for Institutions! Further library commitment to LabArchives. -
• Open Science Framework • Preregistering research • Preprints •
SHARE collaboration -
discussion points • What do you think is the future
of open research and libraries? • What do you want to learn about open research and reproducibility? What are you most curious about? • What are good strategies for engaging departments in discussion on these topics?