 And of course the webinar will be recorded and made available to all of you publicly on the OpenAir website next week. So I was saying that we got a few questions on 90 meter. Yeah, I just went to that but they're not loading for me. Okay, so if you want I can share my screen. Yeah, maybe you share your screen instead. Yeah, while you reply to this and then we see if there's anything else popping up in the chat. Okay, sounds good. Okay, so let me see if you see anything. Okay, so I can see that now. Somebody asking, does my data need to be fair or open? Yeah, that's a good question actually. Yeah, yeah, yeah, completely. So the way I would describe it is that these two concepts overlap. So data could be fair but not open or it could be open but not fair and it could also be both. And I think it depends on the policy requirements that you're responding to whether your data needs to be one or both. If you're responding to the European Commission, they have an open data pilot and they are also pushing for fair data management. That's the way their guidelines have been framed. So really you need to try and address both. But one of the things the EC covered in their guidelines for the pilot was that data needs to be as open as possible as closed as necessary. So it doesn't mean that all data need to be open but ideally you should be making them open if you can. And I think it's better to share under restrictions rather than not at all. So I would see the sharing as a spectrum. Ideally you'd get as close to open as you can. But if there are restrictions that need to be in place, that's not an issue. And similarly, fair is a scale. Well, the perspective we put forward through the Fair Data Expert group is that like a base level of fair would be to have identifiers and metadata and access to either the metadata or the data ideally. But you can make your data increasingly fair by ensuring there's rich description, that it's accessible under common protocols and by making sure that there's a clear usage license so that people know what they can do with your data. So both open and fair are on scales and ideally your data would be both. But there may be reasons why your data can't be open. I'm not sure there are good reasons why your data can't be fair because fair is about making sure your data are reusable. So making sure they're well documented. And I can't think of good reasons why data can't be. Maybe that it would take too many resources or too much effort to make your data reusable. So potentially that's the reason not to be fair. But I think you should be striving to make your data as reusable as possible if you're sharing it. That's a really good answer. Thanks, Sarah. Passing to the next one. Okay. So when you're in a multidisciplinary project with different types of data, how do you create metadata? What format do you use as a software that can automatically generate metadata? So to be honest, you're best to have discipline specific metadata if you can so that the metadata matches your data type and what people from that research area would expect to see described. So the kind of richer information that would be needed for reuse. So it might be that you do adopt multiple different metadata standards in a project that has very different types of data. So some kind of multidisciplinary projects may have some social science data and some genomics data and there are going to be different metadata standards for those data types. So it might be that there's multiple metadata standards in use. There are catalogs of different standards and I'll point to a couple. I can maybe I'll add them into the chat. There's one, the research data alliance metadata standards catalog. And there's another fair sharing which is focused mainly on the life sciences but is doing work in other areas as well now. And there you can search for relevant standards and a number of these standards have tools that help you create the metadata as well. So I'll just search metadata standards catalog. Yeah, and I think we can also add these links to the webinars page as a supplementary resource with the recording. So that whether I ask this question and is not maybe in the webinar right now can find your answer in the recordings and then all the materials as well. Yeah, I'll actually I'll maybe just very briefly share screen. I don't know if that will just boot you off if I just make me stop my screen sharing. Okay. Okay, and then I can just show this resource here. I'm making you presenter again in case. Okay. Here we go. So, so this is the metadata standards catalog. You can see you can view the standards you can search by discipline. You can also view tools that are available to develop metadata for these different standards. And then the other thing the other resource also in the tap is for sharing. And one of the nice things about this resource is it's essentially a series of link databases. So you can look at the standards, but you can also see which databases they're used in and whether the standards are recommended in different policies from journals or funders. Or, you know what domains these standards are covering. So I think that they're two useful resources to look at for metadata standards. Thank you very much, Sarah. Is anyone willing to ask any questions right now from from the participants? Don't be shy. You can you can add things into the chat or we can actually just give you the microphone control if it's easier to just ask a question out rather than typing it. Yeah, we are not going to be quite easy to manage. Yeah. So I'm I'm wondering the two kind of topics that we covered in the presentations were about data management in general. I don't know if you'll have all seen the videos, but then cut did a presentation on reviewing while basics and good research data management and some tips for reviewing. DMPs. So that outlines what research data management is and this need for fair. And essentially what it what it covers is some things that we've highlighted to project officers at the EC when they're reviewing data management plans. So some of the some of the aspects that have the biggest effect on the fairness of the data. So you'll see the in that presentation he gives we talk about the data formats and metadata because that's what makes data reusable if it's in open formats and has rich metadata. We cover data licenses with the need to deposit in repositories because that's quite a big part of the commission's policy and also the need for persistent identifiers. So that's what you'll see covered in the first presentation. And then there's a another presentation where I gave an overview to data management plans and some of the recent trends and the things that we indicated there that more funders are asking for the MPs. I think we've seen a big increase across Europe of national funders starting to introduce policies. Or, you know, enhancing their policies, the commission's policy has changed over the last couple of years. And then there's some shifts in those policies. So I think there's an increasing onus on universities in terms of the implementation of the policies and the DMPs. And also that data management plans are starting to cover more than just data. So looking at the code as well and other research outputs. And then I reflected on how funders are evaluating DMPs and I'm happy to speak about experiences with the commission or other UK funders if that's useful. And then the other trend is just the growing number of data management planning tools and we pointed to a number of those in the presentation. And then the final point was about the kind of future directions in terms of DMPs for open and machine actionable and fair DMPs. And again, there are activities going on through a number of international fora which we can reflect on here in this discussion. Sarah, before I lose it in between the chat, there is a question from Emily about how long you should keep the data as a best practice. Because I don't remember any reference to this in the guidelines from the European Commission. Yeah, so from the European Commission, they don't specify a length, a period that data should be kept for. There's not a kind of preservation requirement per se, but they do say that data have to be deposited in repositories. So they want to make sure that it's in a service where it is going to be more robust rather than just on institutional servers or kept by the researcher themselves and put on their website. It's interesting in the UK, that preservation period doesn't cover it in every funder's policy either. There's a kind of range. Some do specify a period between three, five or 10 years are the most common ones noted. But a lot of them don't specify an exact preservation period. But yeah, I think with the EC, they're really just stressing the repository deposit so that it is in somewhere that's a bit more robust. And I see Sabrina has asked if she could be made presenter so she can ask a question. So maybe if we make, I don't know if you can do that, because I'm not sure I've gone right. And meanwhile, there's another question. What I'm doing this. Yeah, the question from Lovorka Kaya. We should be responsible for the quality of the research data published in repositories because data should be well documented if you want to reuse it. Yeah, so this, I would say this is why it's important if there are subject specific repositories to use those because you're then your data is going to be curated by somebody who is a professional in that given research area. And now repositories won't, you know, automatically check the quality of the data because that's really the researchers responsibility and it's something that we users would flag. But at least I'll be familiar with the standards or if they're, you know, if there were obvious errors with the data or it was just in a state where it's really not reusable at all. I think those things would get picked up a lot more if it's being looked at by people who, you know, work in that particular research field. So I think the responsibility ultimately lies with the researcher. But if you're using a domain specific repository, there are likely to be, you know, some more checks or a better understanding of your data type. Thanks, Sarah. Meanwhile, Sabrina left. So I can't, I hope that she will reconnect to ask a question. I have a new question from Sara Moni. Okay, I think this is in line with the question before from Emily and she's also asking if there's any tip about data management plans tools apart from the NP online. Yeah, so we mentioned a number of tools in the presentation. I don't know if you'll have seen that but there is an increasing number now across Europe. So you've got the NP online from the DCC. There's like a self hosted version of that. Depending on what country you're in there may well be a national kind of version of that. So in Finland there's the Tully service or in Denmark and Belgium and France they have their own national level service. And to be honest, I think you should be using those where there is a local service because there's a group that's coordinating that and they're going to be more familiar with your national funders and you know, potentially they have translations into your own language or a local help desk. That's going to be much more familiar with your context. In Germany, there's a tool called the RDM organizer that's funded by the DFG. So again if you're based in Germany, that's a relevant tool to use. And then the other ones I flagged were some coming out of Norway and also coming out of an open air and UDAC collaboration. The open air and UDAC one is out in beta at the moment and the Norwegian ones I think kind of in the first year or two. So I'm not entirely sure about the status and how much that kind of actively deployed and in use. The other tool that I'd actually forgotten to mention in the slides is one that elixir has developed the data stewardship wizard. So if you're dealing with life science research, that's a relevant service to use. And I think the some of the one of the recent books by Baron Demons, they've been pulling guidance out of that book to feed into the tool. So if you are in the life sciences, that's that's very, very relevant as well. So it really depends on where you're based and your context, what's most relevant to you. Yeah, and assuming that Sarah is Italian, there's nothing in Italy apart from some, well, there's no structure tool as the ones you mentioned. There are some initiatives from single institutions and there's also the translation of the Learn Project DMP checklist, which can be a starting point. And I will put that into the chat as well. Great. Okay. Yeah, so I mean, it's definitely useful to use online tools. It can help. But similarly, if there's just a paper based template that can be useful to researchers. So if there's a template that's translated in Italian, that's a good starting point. If you want to use DMP online, then by all means do it's free for end users. So anyone writing a DMP, but we are introducing a subscription model for people who want to customize the tool. So if you need one, you know, their own branding and their own templates and guidance. Thanks a lot Sarah. Are there any questions? This question is really interesting. Yeah, I'm interested to know where people are from. So are you supporting researchers with data management? Or are you writing DMPs yourself or doing research yourselves? I think you can either type or just switch on your microphone. We are not that many, so it won't be a mess, I'd say. So we have the first answer from Emily. She is a research support person and she's also reviewing DMPs. Yeah, and from University of Ljubljana as well just above, so organizing data support. Sorry, don't worry. So I mean, I suspect most people are probably doing the support side. And I mean, one of the tips I'd give you about supporting data management and DMPs is just trying to have a range of services that you offer. So as I mentioned a moment ago, online tools might be helpful for data management plans but also paper-based templates are really handy. Having like one-on-one consultation clinics so that researchers can come and speak to you about their plan or like a library of example plans. I think you just want to try and have as broad a range of different inputs as possible so that people can then pick and choose what works best for them. And similarly with research data management, you know, there's a whole suite of services you can offer, obviously all of the storage and data analysis support and the preservation. But it might be a lot to address at once if you're just getting started. So working in one area, maybe developing your policy and trying to get buy-in at the institutional level is a good starting point if you don't have anything. So you're starting or doing some training so that you start to get in touch with research communities and start to understand their needs and their kind of priorities for support. Again, another way of getting started. There's a long question. Well, I have no explanation more than a question, yeah, that's true. Okay, so I can read it out. So making research data fair will require that the metadata generated during processing is preserved along with the data itself. The RDM decisions regarding deposit for long-term preservation should consider if this is possible in the repository of choice. While repository and gest will often involve creating new descriptive metadata to accompany the data set, the original metadata should also be accommodated using a wrapper such as Metz as part of ingest. Yeah, and I think this is why, I mean, I've not got anything against generic repository services. I think they're really critical because a lot of research communities don't have a place of deposit. I think this is a reason why if there is a subject specific repository, it's often better to work with them because they, you know, they'll understand what metadata is important to understand the processing steps and how that data is being created and used. And as Garrett said, it's important to keep that metadata as well as just basic discovery metadata. So when you're looking at the fair principles as, you know, four components, obviously the findability, which covers the discovery metadata. But then when you look at the more challenging aspects, the interoperability and reuse, that's, you know, that's all about kind of that subject specific information and making sure that that the data can be understood and are well described. That there's kind of lots of documentation about how the data have been handled and all of the clarifications about any abbreviations in the data. You know, if you've used acronyms to record different values, for example, and information on how things have been measured. So all of that information needs to be packaged up with the data as well. It's not just the data. And there's another question, which is not really the scope of the, well, this webinar, but it will be definitely something to ask during the webinar tomorrow on ask me whatever you want to know about opening her. So if you find, whether ask this question, I will copy this question to mentor to keep attracted for tomorrow. I think that the, well, I'm just guessing the answer. The answer could be no at the moment, but I invite you to join tomorrow's meeting at 3pm to learn more about this. Yeah, I think it would be difficult. I'm trying to think what information would be pulled out of the repositories, but I suspect it would be difficult for open air to distinguish between curated and non curated data sets. Yeah, because that is like a micro data specification somewhere. Yeah, exactly. Open air can read and collect. Yeah. So in, in, in re three, in re three data, you have some basic information about the repository, you know, like what policies it has. I'm not sure if they capture information like, you know, what data process, what processing is done on ingest into a repository and, and to what extent the repository is doing curation, whether they, you know, have a period where they say this data will be preserved. If they, if there's a way for that kind of information to be captured and to be shared, then I'm sure open air could distinguish between, but I'm not sure that that's in any kind of metadata record about repositories just now. Yeah, thank you very much, Sarah. Meanwhile, I posted this question on maintain either. So it would be answered tomorrow. Okay. Any other questions, observations, curiosities doubts. Yeah, it'd be interesting to know what, what things you find most challenging as well, you know, what are the kind of pressing issues for you. So Emily Hermans is, is asked if there's a DMP repository. No, not as far as I'm aware. There are various different kind of publishing venues for, for data management plans. So obviously people deposit in repositories. I know there's a fair number of DMPs from Horizon 2020 in Zenodo. I think because it's a deliverable, a lot of projects will put them in there. One thing that I think would be really great if Zenodo could do this is to actually add a classification type for DMP. So it's a type of record so you can then filter that in the search because at the moment you can't separate our actual data management plans from other outputs that happen to be related to that. In the Rio Journal, that's one of the kind of formal publication venues that accept DMPs and there you can get a list. I think there was about 20 or so when I looked last. Obviously DCC has been keeping a list of DMPs for a number of years and through the DMP online platform you can publish plans. So there's a list there. The other thing that I think is really interesting, Libre have been doing essentially collecting a catalogue of data management plans and they've been doing reviews. So using the Libre community to pick up the strengths and weaknesses of different plans. I think that's a really nice approach so that you're not just seeing a plan. You get some sense of, you know, the value of that plan. And just the final point on this, we actually did make a recommendation in a report we did. I'll post a link to the report through OpenAir and the Fair Data Expert group. We looked at the European Commission approach to Horizon 2020 DMPs and did a survey on that. And one of the recommendations we made was that there's, it doesn't necessarily need to be a separate registry of data management plans but there needs to be some kind of record of them. So it could be within a repository or a separate kind of collection of DMPs. But at the moment they're in various different places. There's a comment from Moitza regarding the challenges and she is saying that the culture of researchers in DMPs and data management in general is a challenge. Yeah, yeah. I think it can be problematic because sometimes, you know, people are pressured in their working lives. There's already a lot of things they're asked to do by the university in terms of reporting or by their funders. And this can just seem like an additional piece of work. And I think it's important that we make sure there are support staff in place to help with this. The initial EOSC report that came out from the commission, I think I can't remember the figures now, but it was a huge amount. It was like half a million or maybe even a significant amount of data stewards that they thought would be needed over the next kind of decade. And it's a big gap because we need to skill people up. Oh, thank you. You know the figure better than me. So I think it is important we have people to support research teams either based within the teams or at a more of a kind of institutional level where people can bring their expertise in. So they might not be able to support full time post within that research group, but they still need access to data stewards or data support in some sense. And it might be a bit of time where it might be some guidance or recommendations to help them set up their procedures, or it might be assistance when it comes to, you know, the deposit process and preserving data. So it could be support that's had a kind of national service level as well. But I think it's really important that there is support in place because you can't just have a policy and expect researchers to respond without there being help. So definitely the culture is a problem. Are there other things that come up as a challenge and I noticed when we asked before about what people do, a few of you had said that you're supporting people and reviewing DMPs. Do you have experiences there to share or questions and comments. And if people want to by all means switch on the microphone if it's easier to just make a comment. No, it doesn't seem that anyone want to start talking apart from you and me. That's fine. And actually just to let you know, I mean, this, this is a kind of a trial thing we're doing with open air. We didn't, we've done lots of webinars over the years and we thought it'd be good to try something different, you know, to make the content available in advance so you can then look at the content. And then come along to the session and just have a Q&A at the session. It might be that that doesn't work as well or maybe these sessions need to be shorter if we run out of questions within, you know, 20, 30 minutes. So do let us know what you think about the format as well because this is a trial that we're doing. Yeah, and something we can adjust for the future webinars or training in general. Yeah. So what I'd suggest is maybe given it a few minutes to see if more questions come in. And if not, we can wrap up. There's one here from Esther. How would you approach senior researchers to set up DMPs? Sorry, I'll just open chat to read the end of that. Even if there's support present, they may be reluctant to change practices. Yeah, I think this can be a really big challenge. Sometimes senior researchers, A, again, they can be very pressured because they're running projects. They've got a lot of demands on them. So they may not have time for the DMP. I think one of the other issues, this is potentially a new area depending on their own skill sets. They may not have come across this before and it can be difficult if something's new, you know, to start learning that or to feel like you don't know what you're doing. So they might not be as open or willing to kind of reach out for support because they feel like they should know in their position. They should already know what they're doing. So sometimes I know a number of projects have often asked somebody else to do the DMP. So maybe one of the more junior researchers is taking control of writing it. And actually something I reflected on in the slides, one of the trends we're seeing in UK universities is that a lot of universities are bringing in requirements for DMPs for PhD progression. So because PhD students are having to write their DMP and essentially update it at the end of the year and talk about how they're managing their data on their project. Senior researchers and like lead PIs are getting familiar with those because they're the ones who have to sign off the DMPs from their PhD students. So that's a way of kind of infiltrating and training without it being direct training for the, you know, for the lead academics. Essentially they're having a role of reviewing the DMPs which is helping to raise awareness of what should be in them and what makes a good DMP and what questions to ask when you're assessing them. So that is potentially a good way as well. And I know when a number of UK unis have done training courses for senior academics, they've often not pitched them as training courses. They've talked about having a research briefing and they've kept something very short. So it's a lot more like a, I don't know, like a senior academic meeting rather than here's what will teach you to do. So I think maybe the pitching of what you're doing is also important to consider other questions. The first one is about how to comply with private data sharing if data needs to be fair and open. Yeah. So one thing I would say about both fair and open is that neither of them kind of contravene or kind of supersede conditions around data privacy. So GDPR, data protection, whatever legislation is in your country, it's still important to make sure that you safeguard participants of your data. So fair and open, as I mentioned before, both scales. Sorry, that's my colleague's phone just going off. I'm just going to mute that. So they're both, they're both scales and you can make your data as open as possible. But if there are sensitivities around the nature of the data and it needs to be kept private or only shared with certain groups, that's not a problem. And I think one of the key things about fair is actually that data can be fair and shared under restrictions. Fair doesn't mean open. So there isn't a conflict at all there with GDPR. But I think with both of them, just bear in mind that you are able to have restrictions on sharing. There are essentially ideals that you're working towards and trying to get the greatest degrees of fairness and openness, but you still need to safeguard participants in your research. And if I can advertise an open air service, Open Air developed a software for data anonymization called Annesia. I'm going to put a link in the chat for you to have a look at it. And if you want to explore its functionalities, you are more than free to do that because it's a free software. Okay, fantastic. There's another question. So in Sweden, funders aren't yet making mandating researchers to set up DMPs. So what would be the suggestion on how to approach them to motivate them? Yeah, yeah. So it's interesting because in some countries, like in Sweden and in Australia, the funder requirements have come later. And that in some senses has provided quite a fertile ground for institutions to define their approach and decide what they want to do. So in Australia, I know some institutions have done more advocacy on the ground and had internal policies that encourage DMPs. And it depends on the institution, what they've done. But it might be that you target a certain group like early career researchers or PhDs rather than having a blanket policy for the whole institution. But I think the key thing when you are in a situation where there aren't requirements is to really focus on the benefits for the individual. That's what DMP should be about anyhow. I think I find it a little bit of a, it can be problematic at times in the UK that we have so many requirements because then researchers feel like they're being bashed from all angles. The university and their funder are saying they have to do this and it can give people quite a bad attitude towards things because it seems like a burden. So I think if you haven't got those requirements, it can be a good opportunity to really focus on the benefits of researchers and to have something that is a lot more research focused. You know, it's actually about what is useful for the researchers to discuss and decide rather than what the funder wants to know to make, you know, certain compliance checks. So you could do more kind of engagement with the research community about what's valuable to them in terms of their data management and to understand issues they've had in the past. And it might be that they've had issues sharing data across their consortium or maybe they've had issues because they're working across multiple countries. And then you can start to pick up on some of those in their data management approach instead. So you're not so driven by some external requirement which might not be a good fit for their needs. And there's another question by Garrett as it is a long one, I'll let you read it through. Yeah, yeah, sure. So Garrett says there's a lot of research data that's been produced prior to the best practice approaches that are fair and open. So good RDM evidence, so example, good RDM evidenced in DMPs. What criteria should we recommend that can be retrospectively applied to make these data sets fairer and more open? So a couple of points I'd make here. I think some data that's already available will have been addressing fairness and openness even before these concepts were kind of coined. So this is one of the points we make in the fair data expert group report that a lot of different research communities have developed practices around how they format and structure and share their data. So essentially they've been adopting fair without it being called fair. So I think there will already be some research that's well structured and reusable and in many cases also openly available. But there will be some legacy data which isn't really fair, isn't very reusable. And I think there's a question about what to do about that. My very first job was working in a repository where I was kind of liaising with researchers bringing materials into the repository. And we kind of focused on rather than doing too much about legacy material, focusing on getting things right from now. And I would err with that approach because there's still a lot of data being created now. And I think trying to improve practice from where we are is probably more achievable than being too retrospective. But we did talk about legacy data in the context of the fair data expert group. And what we suggested there is that that should really be determined by community needs. So if there is a legacy data set which isn't very fair, there should be opportunities for people to make a case for that and say, well this data set could be hugely valuable if it was better described, if it was better structured and to be able to make the case for actually doing that work. But I don't think we should apply any kind of blanket retrospective policies to make legacy data fair because they might not all be that valuable. I think that has to be a judgment about which data sets that's most useful to apply to. But yeah, I'd focus the effort mostly if I was implementing the policy myself at the institution. I'd focus on improving practice from now and potentially doing some work with some legacy data sets, but not all of them. And there are some schemes, so I'm trying to think what they call it, but at TU Delft, they have some, I think it's called a data rescue fund. So they have some resources at the institutional level that can be used to try and improve previous data sets. So something like that could be a good scheme if you did think there's legacy data that's valuable to improve. Thanks a lot Sarah for this. So any, we're quartered to, so any final questions? We actually had a good crop of questions, lots of different things coming in. Yeah. If no other question is coming, I would take advantage of this webinar. And now is another Q&A session that is taking place tomorrow around fair data and trusted repositories with modern growth build from dance at 1pm. Yeah. Yeah, the things on every day ending the information that Sarah already provided you today. Yeah. Yeah, there's a whole program all week of different webinars and Q&A sessions. Some of them will be for more presentations. Some have been pre-recorded. So it would be useful by all means come to other ones, let us know what you think about the sessions, what's worked well and what you'd like to see in the future from them. Yeah, we would like very much to hear from you and have your suggestions. Okay. Thank you very much. We can close it right now and thank you everybody for participating and for asking your questions. And you will find the recordings on the OpenAir website next week. Excellent. Thank you all very much. Thank you. Have a nice rest of the day. Bye-bye. Bye-bye.