 All right. Verla, this will either go really well or spectacularly well. Next up, we have Verla Vanden Heiden. She manages the research data management team at the UK Data Archive, providing expertise, guidance, and training on data management and sharing to researchers across the UK to promote good data practices and optimize data sharing. She also dabbles in various data sharing, metadata, and data infrastructure projects. There's a big push on data sharing. The research fund has updated policies, they found infrastructure, they found support services. So we have a lot of data repositories. A few years ago, the Engineering and Physical Sciences Research Council, this is really easy, a few years ago, the Engineering and Physical Sciences Research Council, with their data policy, placed the responsibility for the preservation, discovery, and access of research data resulting from their grounds of public universities. So universities started setting up their own data repositories as well. So the result is a diverse ecosystem of data repositories. Research data are literally everywhere. We have disciplinary data centers, institutional repositories, international repositories such as GenBank, generic repositories, and also research data in supplements, journals, et cetera. This is fantastic, of course. There's lots of research data out there, lots of repositories. But where do users go when they want to find data? There are repositories, their own search platform, but is it like looking for an evil and a haste that when you as a researcher are trying to find the data that needs? Or is Google the solution to bring this all together? So for example, if as a researcher, I'm interested in finding some data, for example, for some cross-disciplinary work. Where do I go? Do I look at the repository of the University of Newcastle? Do I simply type into Google? Do I scan across all these disciplinary data centers that might exist? I'm talking to people who know about metadata, so you know. Sorry, not too quick here. Just recognize this problem and start the partnership a few years ago. With the Digital Key Racing Center, nine universities, seven data centers from different disciplines. My role in the project is to use those data centers. We have archaeologists, we have material science, we have our environmental data, social science data, et cetera. And of course you know that the solution is harvest metadata, bring it all together. Yes, data centers have different metadata schemas that they use, but these can all be mapped to a central, uniform profile that can make this understanding. So, the consortium was set up, the developer is placed, CGAN was chosen as the software solution, and within a few days we had an alpha version of a portal that was out there. That was the easy part of the work. So here we see it. Next slide. We have a quick overview of all these data sets. In the different repositories. And the tags show you which repository they're held in and other particular characteristics. So now the researchers looking for data on your castle can simply type it into the search box. They accept your new castle and immediately you will see that five data repositories have data on your castle that include data on unemployment, entertainment, angling, early bronze age human remains, overpass, full records, et cetera. That was the easy part, creating that alpha version. It's online. It's been online for months. The challenge then starts. What do we really bring into this data discovery service? Is it data created by UK researchers, of used to UK researchers, data from UK repositories, from public brands? UK DS has got a lot of interesting data from government departments. The archaeologists have masses of data from commercial companies. So what do we bring into this? This, for example, is a really good example. It's data from 30 years of new drugs launched in the UK by pharmaceutical companies. Published in the British Medical Journal, held in the US in the driving post. Should that be in? Should that be out? There's more challenges, of course. What do we understand as a data set? So we have a single image of a superjet postcard in the visual arts data center, survey data, a single crystal structure, an entire experiment run on a neutron beamer, all called data set, by the respective physical data repositories. Even the metadata has got interesting quirks. Crystal structures often have a code as a title. Why is that? Who can understand that? It's something to do with minting the OIS and data site unable to change, put the name in the title once the structure has been made. It can't be done so that name ends up in the subject field. But they have beautiful visuals, which I can't limit here. Also at UKDS we came up with interesting title issues. Lots of data sets with the same title. Fortunately they're old ones, maybe very less. There was a time when series of data sets were given the same title and the detail was in the subtitle. So we need to go and get to the subtitles for some repositories. Licenses. Recent repositories are easy. They've created comments licenses. Very easy categories of how to license. Again with the data centers that have been around for decades, the descriptions of copyright and of license statements that are brought into these facets and categories that are responsible for these services are very difficult to harness. Some people have just thought that the service should only have open data and researchers want to file data, click, download it, use it. That's all we want. All these restrictions, authentication needed, and bar gross. Why would researchers want to have that? But of course we know this is more than that. Understands that we need to have these things. So the solution is we have advisory groups, three, forms throughout the partnership and beyond all these issues, these quirks that we come across, we investigate them, we explain them across the disciplines, we try to come up with solutions. Go at the central service side, solutions at the source with the data centers. And this one example, some of the very loving discussions and comments that go on when the metadata profile, the metadata schema for the discovery service is really defined. So everyone from the partnership can add comments until we come up with a harmonized solution. So where do we go from here? Well, maybe the data palace and during love, hopefully we can match all these data together nicely. The project runs until September, by which time there should be a business plan where you can see how Chisk plans to bring this strong alpha version into a proper service that users can use. To be harvested by another one, that tries to do a global collaboration or something. Yep, yep, yep. Take an 11 OEI out of it together. So most of the hamsters have OEI. Some of it is services for web from the data centers, and then the output will get the OEI. So yes, the idea is, go through that.