Facilitate Open Science Training for European Research Open access and research data management: Horizon 2020 and beyond University College Cork, April 14th & 15th 2015 Using existing institutional repository infrastructure to support RDM David McElroy, UEL • Intro • Horizon 2020 vs RCUK • UEL motivation – RCUK and Internal motivation • UEL response - Research Data Services (RDS) • Workshops & Website • DMP online • Building a data repository • Getting Data • Our first data Developing data.uel Introduction • David McElroy: • Cerif 4 Datasets (C4D) Project Officer at Glasgow • Research Data Officer at University of Glasgow • RDM Officer at University of East London So not much H2020 European experience.. H2020 vs RCUK H2020 RCUK Develop a DMP All but one major funder mandates one at application (EPSRC doesn’t require one but expect ones to be in place) Deposit in a research data repository Some funders have data centres, and expect a deposit (NERC, ESRC) Others stipulate a minimum preservation time (EPSRC – 10 years from last access) Made data accessible, freely to any user Most funders expect data to be freely available (where possible) Provide info on tools/instruments needed to validate results … Source: https://www.fosteropenscience.eu/project/images/presentations/H2020-open-data-pilot.pdf http://www.dcc.ac.uk/resources/policy-and-legal/funders-data-policies UEL motivation - RCUK Funder EPSRC (from 1st of May): Source: http://www.epsrc.ac.uk/files/aboutus/standards/clarificationsofexpectationsresearchdatamanagement/ Key points: • Record all data created • Describe how to access it • Use DOIs UEL motivation – Internal Research Data Service Mandate Library and Learning Services will develop by 1 May 2015 an infrastructure and support service for research data created in consultation with Schools and Services. This will include a portal for datasets which are suitable for sharing. Research data management policy. UEL, 2012 UEL response - Research Data Services • Stephen Grace & David McElroy • What we do: • Workshops & Website • Support (DMPonline) • Repositories (ROAR & data.uel) Workshops & Website • Managing Your Research Data • Writing a Data Management Plan • Sharing and Archiving Your Research Data • Using data.uel to Share Your Research Data • http://find.jorum.ac.uk/collections/rdm Website recently uploaded (thereby hangs a tale...) http://find.jorum.ac.uk/collections/rdm Support (DMPonline) • DCC tool for creating Data Management Plans from templates • Worked with the DCC to build UEL templates DMPonline – UEL PG plan DMPonline – UEL Staff plan Building a Data Repository • Developing data.uel • Early Decisions • Planning & Development • Branding • Timeline & Costs “Research organisations will ensure that EPSRC-funded research data is securely preserved for a minimum of 10 years…” (EPSRC, Expectation VII) Early Decisions • UEL adopted RDM policy March 2012 • Library & Learning Services (LLS) will create a register of datasets… • [and] a portal for datasets which are suitable for sharing • Build on EPrints • CKAN immature, DSpace/Fedora less well supported • Already using EPrints with ROAR • Separate repository to ROAR • Not all data will be open access (ROAR is pure full text) • Workflows differ, presuming researcher deposit • Adapted and simplified ReCollect metadata with DataCite in mind • Development by ULCC (back end) & UEL (presentation design) Planning & Developing data.uel • Functional Specifications • Metadata Schema • Mock-ups • Relational Diagrams Functional Specifications • Excel spreadsheets describing: • What we wanted • Why we wanted it • Who was responsible for doing it • Technologies/plugins we wanted to use • Datacite • ORCiD Leeds have good ideas (which you can basically just copy..) Functional Specifications Description of what we think we need Our reasoning for this. By including this we were able to take advantage of our developers knowledge. If there is a better way of doing something they let us know. Some aspects of development were shared. Above we were to provide the metadata profile Metadata Schemas • Based on ReCollect and Datacite • Only mandatory Datacite fields are mandatory in data.uel “Research organisations will ensure that appropriately structured metadata describing the research data they hold is published and made freely accessible on the internet…” (EPSRC, Expectation V) Metadata Schemas - ReCollect • Created by the UK Data Archive @Essex project • Part of an EPrints plugin which converts EPrints into a data repository • Compliant with Datacite and INSPIRE metadata schemas • Over 40 fields http://bit.ly/ReCollectMeta Metadata Schemas - ReCollect Metadata Schemas - Datacite • Allows creation of permanent identifiers (DOI) • Ireland doesn’t seem to have a member.. • British Library? (DRI are signed up with them anyway) • Metadata Schema • 20 Fields (some sub fields) • 5 Mandatory fields https://schema.datacite.org/ Metadata Schemas - Datacite Metadata Schemas - UEL • Use ReCollect • Mandatory fields match Datacite (not ReCollect) • Added more details (such as ORCiD) Metadata Schemas - UEL Mock-ups • Clear indication of what we want • Red numbers refer to spreadsheet detail Mock-ups • Spreadsheet linked to the Mock-up wireframes • Full descriptions with HTML Relational Diagrams • How projects can be linked to data collections • Potentially to each other over time Branding - Look & Feel • Developed in-house at UEL • (with help from ULCC) • Important to make the repositories feel like part of UEL • Branding • Single Sign On Branding - Look & Feel Branding - Look & Feel Branding - Look & Feel • Modern look and feel • Branding matches with corporate look • Distinct colour schemes for both ROAR and data.uel • Seamless integration between both repositories Timeline Costs Core setup: 3 days EPrints installation, configuration, test repository Phase 1: 7 days Plugin installation & development, metadata Phase 2: 6 days Plugin updates & release, branding, testing Total: 16 days developer time Getting Data – What we offer • Link your publication in ROAR to the data • Archive and share data Collections (data and documentation), managed by LLS • Open access for anyone • Available on application to the data steward • Listed but not available • Description of (funded) Projects with data management plans where possible • Assisted deposit – we’ll come and help you at every stage Our First Data • Large scale survey data • International • Tricky Documentation http://dx.doi.org/10.15123/DATA.4 Our First Data • Large scale survey data • International • Tricky Documentation http://dx.doi.org/10.15123/DATA.4 Our First Data • Large scale survey data • International • Tricky Documentation • What went well? • Very cooperative academics • No rush • What not so well… • Not completely ready to share.. • Complicated consultations http://dx.doi.org/10.15123/DATA.4 Our First Data Our First Data Summary • EPrints can work for data • ULCC are a great software partner (and willing to work outside of the UK) • Clear functional specification/metadata/mock-ups are important if you want a smooth development process • Just because you build a data repository, data is unlikely to overwhelm you in a hurry Thank you David McElroy http://orcid.org/0000-0002-0966-8862 d.mcelroy@uel.ac.uk @davidlmcelroy Research Data Services at UEL Repo data.uel.ac.uk Web www.uel.ac.uk/researchdata/ Blog datamanagementuel.wordpress.com