 Rydym yn ddim o'r fath o'r cyfle o gael lluniau arddangos wrth gwrs o'n ddweud yn eu ddweud. Tre'n mynd i'r byth이 bliadau os yw'r ysgfaith dysgu i ddyn nhw'n ymwinell. Dyna'r ystod y ffordd am rhan o'r ddim enwedd chi'n hyfforddiad. Yr ysgrifennu ei wneud i'w cyfnodd y lluniau gyda'r ysgaflwyntion cyfrydAU i gyd yn ymgyrchu a gyrsanaethau. is really that impact assessment is via publications of citations data only. We think that's narrow in the modern world. We think it's a really limited view. There's information which isn't really meeting the needs of funders and universities. And so there's an opportunity to create something broader, more contextual. And so about four years ago I worked a lot with my team and we came up with a vision for the space and we decided that we wanted to democratise data. We wanted to get data out there so that it was more easy to use. We wanted to de-silo data. And the key thing was that we wanted to contextualise search and connect academics with content. Digital science, as some of you may know, is kind of a schizophrenic organisation. We invest in lots of companies with very pretty logos. And we try to bring them together strategically. So we have lots of people who have deep domain expertise in specific areas. And we can bring all those different domain expertise together to meet needs across the research ecosystem. And so publications are still central to what everybody does as they still remain the atom of research communication. But there are all of these other things that one needs to take into account. And as you can see, we didn't start from a standing start four years ago. We actually had companies who were doing bits of the puzzle already. We had created Grid as a CC0 source, which allowed us to look at institutional names. We had Uber Research looking at a grants database. We had Altmetric trying to pull together the attention from the space. And once we had these ideas together, we wanted to obviously socialise them and make sure that we weren't doing something completely useless for the industry. So we went out and we talked to a lot of institutions and over a hundred different academic institutions and funders from around the world joined us to try and make sure that the use cases we were meeting were really relevant. So we created something that we think is quite innovative. What we did was we moved beyond the publication and we brought in grants data, patents data, clinical trials data into a single database. We also wanted to make things openly and freely available. And so we needed to do that with certain specific types of data. So how do we actually build dimensions? Well, at the core of everything that we do, we need to create a data spine that allows us to hook enhanced data into place. And the core of that spine was actually things like PubMed, Crossref, all of these open data. So dimensions is almost a study in what you can do if you have open data available. And so we brought all of these data sources together to create the core of the dimension system. I40C deserves a massive shout out because none of this would have been possible without the work that they've been doing. And we then were working with actually some very generous publishers. I know that that's a phrase that many of you are often here, but publishers who actually worked with us and actually allowed us access to their full text so that we could data mine it and enhance the metadata that we were making available to academic colleagues. And they did that in a way that allows us licences to be quite free with the data and to allow it to be changed and disseminated around the industry. We worked also with grants. We have about 250 funders in the system. That's about 3.7 million grants. It's about $1.3 trillion worth of grant funding data, awarded data, not opportunities data. And about 300 million of that is grants that are currently held or about to be spent. So it's actually future looking. We have patents data from around the world. We worked with IFI claims. They have 100 million patents in their data system. About 34 million have been integrated into the system so far. So there's a lot of work still to do. Clinical trials. We've been working with all the clinical trials, outlets which make data openly available. And so again, what we've tried to do is pull all these data together, but of course that's only the first problem. The second problem is then to actually enhance it and actually be able to do something interesting with it. So we try to create a modern technology stack. So there's very little human interaction in this. There is a layer of curation in it, manual curation, but almost everything is done algorithmically and using artificial intelligence. And what we do is we take the data from all the sources that we've talked about and we use AI technologies to try and do entity resolution within that set. That then gives us this kind of entity graph of things in how they're connected. So now in dimensions you can look at a publication and you can see the citations between publications, but you can also see the grants on which it relies. You can see the clinical trials which refer to it. You can see the patents which refer to it. And later this year you'll see the policy documents which refer to it, which I think is really exciting when it comes to social sciences. We also wanted to connect academics with content. So we tried to create a way where you could just click a single button in the system and go straight to the best copy, the version of record. So rather than having intermediary steps, there's one click to do that. We also have an API available which allows you to pull data out of the system. We wrote an entire domain-specific language so that you don't need to understand our data structure and you can automatically pull data out of the system using that approach. One of the things that we've done for free, there are two major things we've done for free. One is that there's a badge that you can download from the system. You can take the JavaScript advice that's from the website. You can embed that in your digital repository and you'll automatically get a nice little logo which gives you numbers of citations and some nice little metrics and goes away and gives you a full list of all of the citations. And the second one is that you can go to dimensions.ai right now and you can see the full publications and citations index for free. Anyone in the world can interact with that. Thank you.