Making Annotations FAIR Force2017, October 26, 2017

FAIR Annotations Maryann Martone

The FAIR Guiding Principles for scientific data management and stewardship High level principles to make data: Findable Accessible Interoperable Re-usable Mark D. Wilkinson et al. The FAIR Guiding Principles for scientific data management and stewardship, Scientific Data (2016). DOI: 10.1038/sdata.2016.18

FAIR and annotations? Annotations are data and should be FAIR Annotations make data FAIR by adding searchable metadata and links

Findable F1. (meta)data are assigned a globally unique and persistent identifier F2. data are described with rich metadata F3. metadata clearly and explicitly include the identifier of the data it describes F4. (meta)data are registered or indexed in a searchable resource

Accessible A1. (meta)data are retrievable by their identifier using a standardized communications protocol A1.1 the protocol is open, free, and universally implementable A1.2 the protocol allows for an authentication and authorization procedure, where necessary A2. metadata are accessible, even when the data are no longer available

Interoperable I1. (meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation. I2. (meta)data use vocabularies that follow FAIR principles I3. (meta)data include qualified references to other (meta)data

Re-usable R1. meta(data) are richly described with a plurality of accurate and relevant attributes R1.1. (meta)data are released with a clear and accessible data usage license R1.2. (meta)data are associated with detailed provenance R1.3. (meta)data meet domain-relevant community standards

Kick off Meeting: Force2016 in Portland 70+ attendees came together to: Explore technical opportunities and challenges Explore publisher’s opportunities and challenges Converge on a definition of interoperability Determine use cases that are in scope Identify next steps

Attributes of interoperable annotations Open but standard framework that allows/supports/enables enrichment by community and global discovery to the extent possible. Granular annotations for online elements (html text, images, data, PDF, epub, etc.) Discovery and linking of annotations across different content instances (html vs. pdf) Public, private, group, private group, and authoritative or branded conversations and ability to evolve permissions on selected annotations Selection by both content creators and users Common identification of private conversations Follow/notification Classifications and endorsements, including authoritative endorsement Identities and management of identities among systems Discovery and linking of annotations across multiple versions of content—for scholarly research across multiple repositories, preprints, etc. Persistence as content changes/evolves to new versions. Attributes of interoperable annotations

FAIR annotations: Some considerations F1. (meta)data are assigned a globally unique and persistent identifier: Web annotations are uniquely addressable and they therefore can be issued a GUID such as a DOI. But as annotations are anchored to specific fragments inside of an object, e.g., a span of text, or a part of an image, how are these identified? F2. data are described with rich metadata (defined by R1 below) A critical piece for science is to link annotation capability to standardized and rich metadata via community ontologies and data models. F3. metadata clearly and explicitly include the identifier of the data it describes Annotations include an explicit reference to the DOI and URL of a document, but what about books? Book chapters? F4. (meta)data are registered or indexed in a searchable resource Currently,individual systems provide structured and free text search across all annotations made. What about annotations made by other platforms? https://docs.google.com/document/d/1UObmtnCL_Dw5_tQLJkWgfKBsm_4YpCXxiu3zj2Fsuv8/edit

SciBot: Machine-generated; human curated annotations on the scientific literature Annotation Research Resource Identifiers with additional information Curation private → push to public Based on Hypothesis Annotates to DOI: cross platform annotation RRID → DOI → Cross Ref event database

putting annotation fairness into practice Francesca Di Donato Maryann introduced us to the FAIR principles, (focusing on their declination to annotation). There are many issues to be addressed, and I will start with a couple of examples based on my specific experience with Arts and Humanities scholars.You can consider these examples as use cases. I work very often with teams of scholars. They create digital libraries where they publish critical editions. Their typical workflow is transcribing and encoding texts into TEI-XML and using Pundit to semantically annotate them. 2 recurrent elements I want to stress here: First (not specifically related to annotation): they do not use the same workflow for their work! They compose their articles mainly as doc files, and export them in PDF for the publisher version. Taking no care at all about standards for data fairness and so one. The new technologies do not affect radically the way they work: more specifically, they don’t think about next scholars generations. Write in Word and convert to PDF. Stop. New technologies didn’t affected the way they work. Second: very often, they ask us to have Pundit-made annotations back in the XML source document, again. So, the document is still seen as the “final product”. In such a way, they feel their work is more credited in this way. There are many implications in both their behaviours, but I’d like to stress here is that data fairness concerns not only and not mainly technology and specific actions need to be implemented to change the way science is performed.

Main challenges With this objective the GO FAIR initiative has been launched as a follow-up of the EOSC.

Main challenges “The majority of the challenges to reach a functional European Open Science Cloud are social rather than technical” *Realising the European Open Science Cloud. First report and recommendations of the Commission High Level Expert Group on the European Open Science Cloud

GO FAIR GO FAIR is a bottom-up initiative to start working in a trusted environment where partners can deposit, find, access, exchange and reuse each other’s data, workflow and other research objects. In practice, the GO FAIR implementation approach is based on three interactive processes/pillars:

GO-CHANGE GO-TRAIN GO-BUILD The GO FAIR initiative 3 main processes: GO-CHANGE GO-TRAIN GO-BUILD Education/Training MOOCs, SPOCs Wizards Certification Culture change Open Science promotion Reward systems Technical implementation FAIR data and services Technical infrastructure The first pillar is go change: a cultural change is needed, where open science and the principles of data findability, interoperability, accessibility and reusability are a common way of conducting science. The aim of the second pillar, go train, is to have core certified data experts and to have in each Member State and for each discipline at least one certified institute to support implementation of Data Stewardship per discipline. The last pillar, go build, deals with the need for interoperable and federated data infrastructures and the harmonization of standards, protocols, and services, which enable all researchers to deposit access and analyse scientific data across disciplines.

GO FAIR Implementation Network on annotation Possible actions Communication Advocacy Training Building Hypothes.is, the Coalition and Pundit are supporting the creation of an implementation network on annotation: in practice, it means adopting the go fair approach to establish 3 intertwinning processes: Change, Train and Build. Some actions could include the following activities: Implement Website of the AAKC as a connecting point for federated events and initiatives. Specific local but streamed initiatives of AAKC aiming at: 1) Changing research practices: call to ideas and brainstorming events on: annotation and (labelled) citations / annotation and peer reviewing (hypothes.is experiment) / annotation “status” - specific use cases; videolectures; 2) Training: webinars/MOOCs (such as OpenAire webinars) 3) Annual World event with local initiatives of FAIR annotation? (as in the case of the Open Access week).

Hi! Jennifer Lin, PhD Director of Product Management jlin@crossref.org orcid.org/0000-0002-9680-2328

FAIR is fair. Apply to annotations? Discussion is critical for validation & reproducibility of results Enable tracking of the evolution of scholarly claims through the lineage of expert discussion Make reviews to the full history of the published results transparent (Provide credit to contributors)

Annotations as publication metadata https://support.crossref.org

Publishers: how to deposit Directly into article’s metadata as standard part of content registration process: As part of references AND/OR As part of relations assertions (structured metadata)

But what if publisher is not aware? Event Data collects activities surrounding publications with DOIs (nearly 100mil publications) Annotations are important events! Crossref Event Data event stream contains Hypothesis annotations right now. Interested in integrating more data sources.

Crossref & DataCite APIs Event Data Crossref & DataCite APIs

Making annotations fully FAIR Annotations are important scholarly contributions & can be considered a form of peer review Already being registered by some publishers but support is insufficient NEW content type available dedicated to reviews (including pre- /post-pub annotations). Register annotations (assign DOI) as content in Crossref Through Event Data, track the online activity of annotations as autonomous/independent scholarly objects