 Mae'r ffordd am amlwysgau a'r dwylo gael llwyeddau yn dda, oedd y gyfnodd y web i ddweud i'r ddweud yn dweud rhywunol. Felly mae'n gydag ymddur i'n dweud. Mae'n ddweud i syniadau i'n ddweud ymddur i'r ddweud i'r ddweud am y ddweud o'r ddweud nyfydig o'r ddweud. Felly byddwn i mwyaf i gydag ymddangosol eich bod nhw'n ei ddweud i'r ddweud i'w ddweud i'r ddweud i'r ddweud. a mae gweithio'r dda, ers y gallwn llawer o'r ddweud, yn fawr i gael y cyfnod, yn ychydig yn eich cyhoeddag oesod o ddweud o ddataeth yn ddiddordeb llyfrwng, o'r ddataeth o gyhoeddiaeth i gael y web o'r ddweud o'r ddechrau i ddweud o'r ddweud o'r ddweud. Da oes y gallwn ni'n fawr i dda'i ddweud y ddweud o ddweud, i ddweud o'r ddweud o ddweud, wrth ymlaen i ddweud o ddweud. Ac y cyfnodd cyfu sydd wedi bod cyfeirio i gael cyfnodyniadau cynserau yn ddigonol wedi ddegwyddarhau. Yn y dweud erbyn gyda'r cyfnodd iawn maen nhw'n ddefnyddiad ym mhawr hwnnw a'r cyflodydau cyffinodd yma hyn yn dweud, ac eftanol Ewfio cyffinodd o'r cyflodydau cyffinodd o gyflodydau cyffinodd o'r cyflodyd acau'r cyflodydau cyffinodd o'r cyflodydau cyffinodd. Ond, ydy'r clubs ei physio o'r ffordd, o'r ffordd, o'r busio cyfrannu talog, o'r bwrdd, o'r fforddau lliar. Yn gyfnod i'n cael ei phirio, mae byddai'n eistedd ymlaes i gŷtio'n du o'r ffordd i'r gweithio yma. Dyma'r dwi'n gwheil ychydig oedd gwahanol i ddechrau i'r newid yma. Yn cyfnod i ddegwyd ar gyfan, rydyn ni'n fwylo, y cyffredin ni'n rhanol i ddweud bod Nigel mae'r natur o'r holl hwn yn gweld o fewn hyfforddiad, o'n dweud o'r hanfforddiad ar y gof rebell surrender a'n celf a'r holl sy'n gweithio wahanol. Mae'n dweud oherwydd maen nhw'n cael unrhyw hwn o'r gweithas yng Ngheilwyr'r Ddes Gweithléol o'r oeddod o fi'n meddwl hwnnw a chi'w Gweithlion Dysgrifennol o'r pryd yn gweithio ddim. Felly, dwi'n rhoi'n credu'r ffasau o'r hyn o'r hynod oherwydd y cynddo is that there are an uncountable number of different kinds of impacts that you might want to create and all of the interesting and important problems are very complicated and generally require data from lots of different sources collected in different ways. So I think the way that we're trying to think about this challenge is that before someone creates a positive impact, so deciding what's the best medical treatment for some kind of condition or what's the best way to run our education system. Before you get to that point, that decision maker has to be informed by some complex analysis by a skilled data analyst. That data analyst needs to get hold of their data and to understand what it means. That data needs to have been collected and processed and managed by someone. So we're kind of going upstream in what is a fairly complicated and supply chain, a kind of value chain, if you like, that has lots of steps and you only get to the point of creating some impact at the end of it if all of the steps in that value chain exist and are well supported. So I think where we face big challenges at the moment, we've got massive opportunities but also we're not achieving those opportunities as quickly as we might. So one is just that there aren't enough good data analysts around. So it was great to hear what Jill was just saying about the work of the data lab to try and almost creating a new profession of data scientist that is bringing together some traditional data analysis skills with working with different kinds of data and different ways of delivering data to meet some of these problems. And we definitely need a lot more of them. We see that in our government work that there's a lot of data there but it could be used much more effectively if more people working in those organisations had the right skills to exploit it. Some of it's just an awareness problem. They don't know what's possible but some of it is that there's only a small number of people that do it and the good ones are always vastly overworked and underfunded and so on. So increasing the size of that profession is a very important thing. But I think the big obstacle for us is that this is really still a cottage industry that the analysis work that goes into supporting these kind of decisions is that you're dealing with generally highly skilled analysts but if they were cabinet makers they'd still be going out and chopping down their own trees. Anyone who's actually tried to do this sort of data analysis work know that you'd spend, I think somebody's quoted a figure of 80% of your time gathering and cleaning the data. And I think our value chain here is crying out for some industrialisation. We need some standardisation, kind of componentisation, enable things to be automated so that the work of the data analyst can be that much more efficient because we have say thousands of different kinds of problems that need to be solved and we don't want each of those people to go out into the forest to start chopping down their own trees. The upstream parts of this process can be improved. And so one of the projects that we've done is the one that Roger mentioned. We've worked with his department and the Scottish Government on the statistics.gov.scot site. And I think one of the reasons that I think that has gone very well and relatively smoothly is that Roger's group has responsibility for a wide range of different topics of data. And so we're able to apply a bit of central standardisation and good practices to data about health and crime and education and environment and so on. And therefore come up with a data resource which is kind of well integrated with itself, if you like, and covers a broad range of things and therefore has a lot of standardisation built into it and so can support effective use of that in lots of different ways and effective combining of data about different topics in a single way. So that's all very well but say if you look at the kind of broader picture of government where what if you want to combine that data from Roger's group with stuff from DWP or from local authorities or from businesses. So there's a big standardisation challenge there and the lack of that standardisation I think is the thing that gives data analysts who should be doing highly skilled work and awful lot of donkey work that they have to do before they can get there. So what I think we should be doing to try to increase that impact is to look at each step in that value chain and think what can we do to move to a kind of more mature model of delivering all of that stuff. And we see ourselves as sitting somewhere in the middle of that. We're providing tools that help the owner or manager of some collection of data to curate it and deliver it to people in ways that save time for the analysts. But our bit of the puzzle can only work if somebody else is collecting data and obviously there are all kinds of new ways of collecting data available now. Someone has to collect that and to process that and to document it and to deliver it in a way that is amenable to use. And similarly our bit of the puzzle is of no use whatsoever unless there are data analysts out there who can take that data and apply it to a problem. And so part of our work is to try to make life easy for data analysts. So I suppose kind of what I'm saying is that I think we should be trying to do and groups like this can hopefully contribute to that is to think about how these different roles in the final delivery of impact together and try to move towards less of a cottage industry and more of a sort of industrial revolution of creating impact from data. Thank you.