 Ddau'r gwrth ymyrch. Felly, rydyn ni'n gwybod y pethau'r gynlluniaid, rydyn ni'n mynd i'ch fflii'r cyfreith. Felly, yn rhan i'r ffordd, mae'r gweithio'r pethau yn ddifthio'r gwybod yma. Pau, rydyn ni'n gwybod y pethau yn gweithio'r gwybod. Felly, ddod, ddod, ddod i'n gweithio'r gweithio'r gweithio'r gwybod yma. Mae'r rhaid i'r gweld yn rhan oedd Giscid yn rhoi'r ddaeth yn gwybod, mae'n gweithio anelitogol iawn, ymgyrchiau cyfnodd, rhaid i'r programu data a'r reddoi'r archifau, rhaid i'r bobl yn ddoddoddodd yn ddigon i'r gwsgwllen a'r ysgrifennu challenges on the manager of the research data program. So, there will be some reiteration of some key points from Graham's presentation, but I hope that is all to the good. So, it just considers it a priority to support universities improving the way research data is managed and where appropriate made available for reuse. The where appropriate is important of course. There are various drivers for this, including thunder policies. Not a lot has been said today about the research council's policies, but they and the recent, now a year old, EPSRC fund research policy has caught the attention of a number of institutions as it requires a data policy, a roadmap for provision of a data support service over the next three years, and indeed a catalogue of research data holdings. Legislative framework Mae'n gweithio'r byffordd ar yr awdraeth, faintiaeth, iawn gwneud panch, yn ddodolion字 mewn gweld, ac mae'i rhai ates. Mae'n gweithio'r byffordd yw'n ymwneud, mae'n amrywod yn anghybodaeth am gandol iawn, ac mae'n gweithio'r gymdeithason gyda sefydliadau sydd gennu yn gweld ddiogel eu cyfennidegau arddugol. Rwy'n gweithio'r cyfennidig ein ddechrau erbyn, arwein ddych chi'n cael ei fydd iddo. Ar ales o'r strategiaeth ysafodol ei ddatgu'r datgu. Gall y cyfleu gwneud yn gwahanol talog gallwch chi fod yn cael ei bwysig yn deilabol, yna sy'n hollwch i'r hawddyn yn gweinol, a oes i chi fyddech chi'n gwahanol a'u rhan o'r rhan o'r halwch oedd sicrhau'r data arall yn dod o'r perthynau o'r homeon sydd chi'n bobl i'r yrhwyro yng nghylch i'w rhanion. Mae'r bobl ei sydd yn gwahanol i'r data arall yn rhanol, ac mae'r bobl eich bobl eich bobl i'r data yn eich bobl i'r bobl i'r bobl i'r bobl i'r bobl, ac mae'r ddweud yn gwam kommtr uchydig, Alec. Mae'r sefydliadau bwysig o ein hirionedd rhoi'r perlwyddiad. Felly y mae o'r linell yn gallu gŵith iawn o un Powh. Mae'r byeun yn gwahanol hyn sy'n meddwl, fel hynny'n gallu gwahanol ar hyn o'r busgau 1112. Mae'r busgau'n mynd o'r busgau yn mynd i chi fod yn fawr o'r mewn cyfasig fel y Fawr mewn busgau 1112. Mae'r busgau eich hirionedd o'r busgau, yn fanyddig ar gyfer busgau mewn prifodus. Rydyn ni'n gweithio i'ch bod gennychwyd yn gweld'r pobud with the institutions in which we have projects. Ysgrifennu ar holl gwybod rysgrifennu hon i'r ffordd yw i'ch gweithio'r unig yn y Tu Rusull Group yn y 94 Group University. Mae'r ddefnyddion perthynau gwaith rai i'r ffordd hynny wedi gwneud hynny, ond rydyn ni'n ddiweddio ond mae'r cyffredig anonymidi. Yn y Mizgrifennu, y Tu Rusull Group Unwespledig wedi bod y dynnw'n rhaid, hordings nifer hwydd o pethau papyrd. Mae cyflod o bethau ddysgu a ddysgu hwnnw ymgledd. Mae'n ystyried o'r unrhyw ar ddannig, mae'r 800 terbyddol ddau gweld y ddweud cyflod yn ddweud, ddaodd ddweud 300 terbyddol. Neu sayu ddweud o ddysgu hwnnw wedi'u ddysgu, when institution gave the estimate that the 800 terabyte store will be full in the next 12 or so months. The point of this is that there is significant amount of data in temporary storage in external drives etc. and an email I received this morning which was quickly incorporated into that slide The more groups we go to talk to the more we're hearing of significant data holdings on external hard drive system etc. I think it is fair to say that a lot of universities really don't have a very accurate idea of the sheer quantity of research data which is being held. It doesn't mean that all that research data needs to be made available for reuse. It doesn't all, as we heard from the science institute this morning, doesn't all have a reuse value, but a lot of it will do and it's important to know what you have, where it is and what the potentialities are. Similarly, the 94 group university doesn't provide any central data storage provision that's devolved to the faculties, but I think you can see from that slide there's all significant data storage issues. They're notably in engineering more than the other faculties, a 170 terabyte faculty system with an urgent need for that to be expanded. And in recent interviews they've met one group which currently has 250 terabytes of video data as it happens, only half of which is in managed storage, the rest is on external hard drives etc. This is a common place and it's potentially an issue because of course there's a risk with the sort of unmanaged storage. The data centre under the debt which comes with attendant risks of data loss. This has appeared on a couple of blog posts. If anyone has seen this absolutely heart-rending story of a PhD student losing an external hard drive with five years of research data on it, this is not the sort of thing that we want key research departments to be doing. However, it's not just about storage or avoiding data loss. It's about good research management and good research management as we've heard is about the triage and knowing precisely what to keep and what to throw away. And I don't think this can be stressed enough. Graham's made the case very, very strongly. It's from the researcher perspective as well about making the most out of the data created. Now I think the Australian National Data Service that's doing a lot of work in this area has a nice take on what that requires, what it is that's required with the actions, the support etc. that's required to make the data more reusable and to turn from disparate data objects if you like into structured collections and data that's from being unmanaged to being managed, disconnected to being connected, invisible to being findable, findable is extremely important. And from having a single use to being reusable. I think those principles that mantra if you like is a very elegant way of stating a lot of the activities which the managing research data programme and which the digital curation centre is involved with. So in the remaining time I'm going to give you a very quick sort of three-part history of what we've been doing in the managing research data programme in related activities to tackle the research data problem. Prior to 2007 up to 2009 it was very much a matter of research understanding the problem that was done through a series of reports which I'll go into very briefly. Then with the first managing research data programme very much a character of prototyping solutions which since the second programme got started in 2011 there's been a process of hardening those solutions and above all an emphasis on building capacity and hopefully at the end I'll have a couple of minutes to say what the next steps are. So important, just funded reports and other work in the area of understanding the problem. What is the challenge relating to research data in institutions? Three key reports there, there were others. I'll just mention as the role of data scientists and curators we've mentioned earlier an important report looking at those career structures and the recommendations of which we're trying to implement through training activities within the managing research data programme but obviously there's a broader activity within the whole sector required to build up those skills as been stressed. The first managing research data programme got started in 2009 and was designed on the basis of those reports and particularly the fourth report that I didn't mention, the UK RDS, UK research data service scoping study which did a lot of useful information about the state of data curation in universities. So we've designed it along four strands, a set of projects looking at infrastructure in the broadest sense in institutions both in terms of the human infrastructure, the support infrastructure and the systems themselves. At this time there was a, and still is to a large extent, an important focus on the creation of data management plans as a way of inculcating good practice. So we explored what the challenges were with this, how to make them usable, how to meet particular disciplinary challenges. We also had projects producing training materials and projects looking at the challenges of data citation and data publication and this is the important downstream area if you like. If the research data can't be cited, if it can't be reused, then there's little incentive for the researcher to make the research data available. The second programme carried basically the same, sorry, I'm just going to go back at a slide. I should have mentioned that there's a lot of out, we funded a lot of projects, they produced a lot of software, guidance material, training materials, et cetera. The address for the outputs page there is well worth a look. There's a wealth of material there which I'd like to draw your attention to. The second programme followed basically the same structure, as I've said, with a far greater emphasis on building institutional capacity. So there's 17 large projects developing research data management policies, developing the human infrastructure and the technical systems. We continue the emphasis upon exploring what it is to produce the role of data management planning in inculcating good practice, what a good data management plan looks like, et cetera. Early next week we'll announce a new set of projects to produce training materials and projects to develop innovative data publications and I said that's the incentive side, linking research data to publications which extract knowledge and incorporating the publication of research data into the scholarly life cycle, wherein we can get reward and recognition for researchers who do share their data. So we like to think that in these programmes we've followed both in the programme and the projects we've followed, a holistic approach which has joined these five areas. The necessity within institutions for leadership and policy development, the need for guidance and training for researchers and for research support staff, support for data management planning if I stress, the development of prototype systems and infrastructure which has been hardened by further projects and as I've mentioned the publication, citation and discovery mechanisms. Some of this is being reported on and synthesised by the Digital Curation Centre who have in development how to guide, how to develop a research data management service in an institution and that will serve as an introduction to a larger toolkit and that strikes many of the same target areas that I've mentioned. So finally, next steps, we've done a lot of work on building capacity in institutions as I've stressed. The next step we feel is to try and build some building blocks of a national infrastructure and information systems. So in particular, as journals are increasingly implementing policies that are requiring the availability of research data, these are proliferating and it would be a good thing if researchers and research administrators and librarians had easy access and reference to those policies and their contents and so we'll be funding from June a feasibility study rather for that sort of service. I think it's also important now that universities are developing catalogs of their research data assets to pull these together in a registry of research data to facilitate discovery and encourage reuse. Again, with a mention of ANS, there's a good example of that in the Australian National Data Services Research Data Australia Discovery Portal and I think that's something which would benefit the higher education sector in this country as well. Link's there for further information and I'll be keen to answer any questions later on. Okay, we have a little bit of time for questions. Any questions for any of our lightning speakers? You're feeling worn out. We have a question down the front there. So if I share my data and some random person picks it up and uses it, that's great. If they make some discovery that's a complete junk, how do you avoid something like a secal hoax where someone picks up some say particle physics data, aligns it with the phases of the moon and the sweet streams or whatever and ends up saying, ah yeah, it's bosons caused wars or whatever. How do you guys see peer review of data usage I suppose is the question from the left? Okay, who's left? Peer review of data is an interesting question. There are a number of data-focused journals increasingly that are exploring, that already have some guidelines on peer review of data. I think it's still a question that requires attention. I think the most important thing is to have the data available so it can be community reviewed. I'm talking more about the use of the data. So if I have a data set that's publicly available on whatever and you use it wrongly because you don't understand maybe the process that I've gone through to create that data or you've misread the metadata or you've divided by zero or something like that. Wasn't it always thus? Someone does some bad science. You point this out. You misuse the data. Your use of the data is wrong or flawed or... I think this is just all incredibly tractable within scientific disciplines. So if the volume of data is exabytes, which is not far off, that's a huge amount of data that I can do all sorts of wacky crazy stuff with. Tracking that provenance through those data sets becomes increasingly difficult as well because you end up tracking more provenance data than data itself. Tracking provenance, again, that requires good quality data. Again, I don't know. I think I've sort of referred to my previous answer. So you can see in the climate science world the discussion around models and it's incredibly complex science and in some instances it's very hard to communicate publicly and very hard to defend from attack both within the science and outside it. But I think that's just the world we're in and I'm not sure that there's a snappy answer and I've taken up too much time anyway. I think given the time and given that felt like a good discussion to have during the drinks reception, I'm going to move us on and wrap up this little bit of the session and move on to our next speaker. So can we just say thank you to the three lightning speakers?