 Hi I'm Sarah Jones from the Digital Curation Centre and as part of this foster and open-air webinar series for Open Access Week I'm going to be speaking to you about data management plans and what you need to know about DMPs. This webinar is aimed at those who are supporting data management plans and I'm going to pick up on some of the recent trends and developments in the area. So first off as I'm sure many of you are aware more funders are asking for data management plans and the requirements from funders are changing. There's an increase in the number of requirements across Europe. I expect you're familiar with the European Commission's Open Data Pilot and the ERC requirement for DMPs but we're seeing a number of national funders also asking for data management plans like ZonMW in the Netherlands, DFG, the Academy of Finland, the Research Council of Norway and the Swiss National Science Foundation. Most of the research councils in the UK also ask for data management plans. Now this list here isn't comprehensive there are bound to be other funders across Europe that I'm not aware of that are starting to bring in requirements but what we're seeing is just a growing trend towards more DMP expectations and the way those those requirements are being formed is also changing slightly. Traditionally data management plans were always asked for at the application stage but increasingly we're seeing funders asking for data management plans during projects or they're noting that they should be living documents something that's updated during the course of research. NERC in the UK has an outline data management plan at the application stage and then they expect a full plan to be written in conjunction with their data centres once a project has been funded. Other funders may expect the DMP to just be produced post-award once the project has started such as the European Commission, the EPSRC or the Research Council of Norway. With the European Commission the first version is due as a deliverable by month six but they expect that to be updated whenever changes occur and particularly at minimum at the review points for the project. We're also seeing an owners being placed on universities for the implementation of the data management plan. In the UK the EPSRC brought out quite a groundbreaking policy in 2011 which articulated the responsibility of research organisations to support data management so universities should have appropriate services and policies in place to make sure data are properly stored and managed and shared and the EPSRC also say that universities should ensure the data management plans are created and maintained. They expect them to be in place but they don't necessarily want them to be submitted to them. The Arts and Humanities Research Council in the UK updated its policy earlier this year and they now require that people submit their applications confirming a number of statements and I've given an example of one of the statements here. It says that the institution is able to store the data appropriately and that the relevant people have been consulted so really they're expecting that the university has internal processes in place to ensure the data management plan is created properly and that that is implemented during the course of the research and most recently I heard that the Research Council of Norway is introducing a policy which expects the data management plans to be approved by the university before submission and also that they'll be made available openly online. We're also seeing that data management plans need to cover more than just the data. The Wellcome Trust issued new guidelines in 2017 asking for an outputs management plan which covers the data sets, the software, new materials like antibodies and cell lines and also the IEP and the EPSRC also has a requirement for software management plans because code is often created on their research projects so when we're thinking about data management plans I think we need to consider data in a very broad sense to cover all of the outputs of the research project that are important to validate that research. In the UK we're also seeing that universities are starting to require data management plans for PhD progression and the idea here is really to start people off early and to build good practice from that first stage of their research careers. Supervisors are often expected to review the data management plan and approve it so that people can progress to the next year of their PhD and this is seen as a good way to get the academic community at large more aware of good practice in data management and sharing and there's a discussion about this on the Data Man lists serve recently and I've highlighted the universities here that have spoken about their policy and their expectations for PhD students. It's not a UK phenomenon though, TU Delft is also introducing a requirement for PhD students from 2019. We're also asked quite regularly about how funders are evaluating data management plans and a number are doing so although sanctions are still quite rare. I want to reflect here on the European commission's approach to reviewing DMPs under horizon 2020. As it's a deliverable it's usually the project officer that looks at and assesses the DMP though we know some of them do ask the project reviewers or external experts to assess the plan. Regardless of who actually assesses it, the assessments made based on an internal framework the commission has which is aligned with the fair data management template so it takes the questions within the template and it asks the reviewer to say whether something's been addressed fully or partially or not at all and to make recommendations for changes. We've done a whole series of training courses through the Foster Open Science project where we've been working with project officers on how to review DMPs and there are a number of aspects we highlighted when they're doing their reviews. So we picked up on the data formats ensuring there's a good description of the data that ideally the data is in an open format or a format that's in common use in that research area. We really stress the importance of metadata because that's central to fair so there should be a rich description of the data, use of domain metadata standards and controlled vocabularies. We mentioned the importance of repositories and depositing ideally in domain repositories if they're available or generic services if not and we also touch on the importance of PIDs and that these may be assigned by repositories and licenses and that they should preferably be open or if there are restrictions non-commercial or no derivative licenses that these are justified. A number of funders have guidelines for reviewers on assessing data management plans. The AHRC asks this series of questions so they're looking at whether the information in the DMP is appropriate, whether the plan seems feasible and valid overall. The ESRC, the Economic and Social Research Council, has a whole document of guidelines for peer reviewers and this can be a really useful resource if you're looking at assessing DMPs too and in the UK the RDM support community has been developing a series of rubrics based on the different funder templates and you can see a number of these on the research data network community collection on Zonodo. As I mentioned, sanctions are actually quite rare. The DMP is often considered separately to the scientific excellence of the research proposal so a poor DMP could be a reason for rejecting a proposal but usually conditional awards are made so projects may be asked to reconsider and update the information in the DMP instead. That said, there are a couple of cases where there may be sanctions. The ESRC do note that the final grant payment could be withheld if data aren't offered for deposit and we do have anecdotal evidence from some universities in the UK that a poor DMP can be the difference between funding or not so if you've got two very strong proposals where the scientific excellence is really high a poor DMP may just let one down and that could be the difference. The other area where we've really seen trends emerging is the amount of DMP tools that are becoming available. The DCC offers DMP online and we've been running this service since 2010. You can see the the homepage here within DMP online there are a number of templates for different research funders from the UK but some Dutch funders the EC and the ERC too and universities can customize the tool to add their own templates and guidance or example answers and you can also review usage statistics for your organization. One of the new features we've added is plan review so you can enable plan review and give feedback to users on their DMPs if that's something you want to support and researchers are also able to share their plan they can co-author it with others they can choose to make it visible within their university as a reference for others or they can make it public and they can publish it via the platform. You can see here what it looks like when somebody's writing their DMP you've got the different sections of the the funder template this is a horizon 2021 and the the answers being included here there's also guidance from different sources from the funder or the DCC or your university and this allows researchers to drill in to support depending on what level of support they need so you can turn the guidance on and off. DMP online is based on an open source code base called DMP Roadmap and because of this we've seen a lot of other instances of the tool emerging in different countries so in Finland there's the DMP Toolie service in Denmark there's a service run by Dyke there's another service in Belgium and in France there's a tool called DMP Oppador so depending on what country you're based in there may be a national level DMP service that you can use. In Germany there's a tool called the RDM Organiser which is funded by the DFG and this is a slightly different service model it's an open source code base like DMP online but there isn't a central instance run at a kind of German national level the intention is that different universities take the code and they deploy their own local instance. There is a demo version available low at this URL so you can log on there and you can see how the tool works. There are a couple of tools emerging in Norway one called EZDMP which is run by Sigma in collaboration with UDAT and one by the Norwegian social science data archive the NSD and I hear there's plans to have a kind of single interface that will lead into these two tools so that researchers in the Norwegian context just have one place to go and then they can adopt whichever tool is most appropriate for them. OpenAir has also been developing a tool with UDAT called Open DMP and this is the main difference here really is that they're shifting from funded templates to the data profile as a central entity within the tool. It's currently a proprietary format but they do plan to adopt the RDA recommendation once that emerges so that the DMPs are machine readable and the code is available under an all rights reserved license it's on the UDAT GitLab repository. At the moment the tool is in beta you can get it at this URL but I think once this becomes active as a service the intention is to make sure that both the DMPs and the code base is open. So you can see there's a growing number of tools both across Europe but also internationally and we've been tracking these on a website called Active DMPs. We'd encourage everyone to update this if there are other services you know about or other requirements that are emerging. The content for this website is actually held within GitHub so you can submit a pull request to update details on the site or you can email and drop us a line and we can update it for you. So finally I want to touch on some of the future directions in terms of DMPs and the idea is to make these open and machine actionable and fair and there are a number of international fora working on this through the RDA there are two working groups one on exposing DMPs which is developing use cases to make DMPs open or elements of them open so that that information can be reused in a number of different contexts and one on developing common standards and the intention here is to develop a common information model so that data management plans can be machine actionable and hopefully once this model emerges different tools will adopt that. There's also a working group through force 11 on fair DMPs and I wanted to pick up on this notion of fair DMPs first we know that some templates cover fair so for example from the European Commission and I think this can be a really useful way to ensure that researchers consider data access and reuse and make sure their data are fair but the concepts within fair can overlap and this can cause repetition and annoyance if the template itself is structured according to fair. You can see a couple of comments here that we received when we did a survey on the European Commission approach to data management plans so I think it's important that DMPs consider fair but not necessarily that they're structured in that way. I think a more useful way to think about data management plans and fair is that the DMP itself is a fair object that it's published or deposited in a repository that we assign persistent identifiers and metadata to the DMP and that we follow standards for data management plans as they emerge and this will enable us to reuse the information within the DMP and this notion was put forward in the Fair Data Expert Group report there's a recommendation that DMPs act as a hub of information on the fair data objects and that they connect together the wider ecosystem so the DMP is connecting with repositories and metadata standards and different registries and this is something that we're working towards with the DMP roadmap platform to ensure that the information in DMPs can be reused. So the final question I wanted to ask is whether DMPs should be open and in many cases they can be and projects are willing to make them open when we did that survey with people who write DMPs for Horizon 2020 almost half said they would openly publish their DMP but there are various cases where that's not applicable so if a project hasn't been funded yet or if the research is sensitive and information in the DMP would expose the location of animals under threat of extinction or whatever the research is about it might not be appropriate to share the DMP or it might only be appropriate to share certain elements of it but I think we should be encouraging researchers to make their DMPs open because sharing early and openly enables us to reuse that information both to inform service delivery but also to help data to be discovered and reused and we know that when DMPs are made available others find those examples really useful to to enhance their own practice and learn what good practice in their research field is so in open access week I think we should be encouraging researchers to make their DMPs open so that's the the main aspects I wanted to cover in this webinar thank you very much for listening