 It's just as sealing into the environment agency and I'll be coming over and just Miss acabar that it is speaking about the lecture that has started only on Friday. So I hope you haven't been thinking about the environment agency�ה ty wrong and that has been done to the other issue. Thank you. You haven't heard me talk yet. Sir, thank you very much for letting me go here and talk to you about our open data journey. I expect, over the course of the next 20 minutes or so, Mae'n dda'n meddwl yn ei wneud i gael, ond rwy'n meddwl i'r hunain. Rwy'n meddwl i'r hyn. A fyddai'n meddwl, oherwydd, y gael y gael y gael y gael y gael y gael y gael y gael, ond rwy'n meddwl i'r hyn yn ei ddim. Mae hwn yma yma wedi gweld i'r hynny, i'w wneud, a'r hyn yn ei ddim. Y gael y gael y gael y gael y gael, i'w gael y gael y gael y gael, yn yng nghymru i gyngorodd ac yn ymwneud i'r hynna. Felly, mae gennym o'r ystyried o enghreiffton, ac mae gennym o'r perffeydd yn cenderfynu'r gyfyrdd rwythbeth. Felly, ond y cyfryd ymgyrch yn ei ddweud. Felly, mae gennym 10600 o'r ymdweud, ac mae gennym o'r ymdweud. Mae gennym o'r llwybr, o'r llythbeth yn Brystel, o'r llythbeth yn London, ac mae gennym o'r llythbeth o'r llythbeth allan i'r llythbeth. Mae'n gwybod yw tu wedi cael ei meddwl i'r ymwysig a'r ymdwynd. We look after water quality in rivers and we look after water resources, so abstractions. We do things like drought. We also look after fisheries and you may come across us looking after and being involved in flooding. All of that activity is underpinned by data and information. That's what powers our organisation and helps us make the decisions we need to make. So, who am I and why am I here? I've been in the Environment Agency since 1998 and I know I don't look old enough. It's the water quality that keeps me looking young. I started off throwing buckets in rivers and wandering around landfill sites and collecting data from the environment. I moved through to analysing it and using it to do reporting and targeting our activities. More recently, well, since 2001, I've been involved in sharing our data with third parties. So, that's actively going out and working with other people for them to use our data and information. I manage a team, so Open Data Manager is kind of like my new job title. The team that I manage within the organisation is the team that has been responsible for commercialising environment agency data, so making revenue from the environment agency data and information. I thought this wasn't going to work. So, I know you can't read this one because when I tested it on here, it didn't come up. So, what I'm going to take you through over the next little bit of time is three sort of three bits on our Open Data journey. I've kind of added the third one very recently in light of some of the stuff that's been in the press over the last week that you might have come across. So, what I've called our Open Data incident, how we've made decisions around Open Data and then what the future looks like for us. I should have said it earlier on. If any of you have got any questions or something I say is not clear, just shout out at any point. So, in the past, well, since 1997 actually, we've been sharing our data and information from a central point widely for people to incorporate into products and services. That was in response to companies coming into our local offices and effectively photocopying records and then digitising them themselves and creating products off the back of it. We should be doing this ourselves. One of the fundamental premises that we were doing then is that we should recover some revenue from that. We should charge people for accessing and using our information to make their environmental decisions. It fitted with the government policy at the time, so we've basically been charging for data since 1997. Initially, it was everyone, so public, private, commercial, non-commercial, we would charge. The only exception was if it was a very, very, very small bit of data for your individual locality, so a national dataset we would charge for. In 2010, we changed that to only charging people for the commercial use of our data. So we now receive about £6 million worth of revenue per year from people using our data in products and services, predominantly conveyancing. So if you do a property search report or get a property search report when you're buying and selling a house invariably it will have our data in it and we get paid a royalty for that. Alongside that, and this is the balancing bit, we've been exploring open data. What I mean by that is we've been releasing some of our less valuable in terms of we didn't think we could get a revenue from it data as open data for a reasonable amount of time. And with one notable exception, which is the bathing water data, which we've released I think now for five years or so as linked open data. So there's products and services built on that and from the start that was designed to be free. So we've been kind of dabbling, but very much focused on making some revenue. One of the reasons, well one of the other reasons that we went down that route is for us the contractual relationship that we had with the people using our data was really important. We were worried that people, if they took our data just incorporated it and we had no control over what they were doing, they would use the wrong data, the wrong information. It wouldn't be shown in a way that was appropriate to what they'd been given. So that kind of one-to-one relationship with the people using our data was important. And all of that approach was in line with a piece of government policy called the wider markets initiative which encouraged government bodies and departments to sell spare irreducible capacity for us data to make that revenue to support extra activities. So we weren't doing anything wrong. However, you may remember Christmas 2013 to May time 2014 we had a period of unusual weather activity. It started in December with a storm surge which was a body of water coming down the North Sea funneling into the bottom end of the North Sea in the Thames Estuary which was bigger than the 1953 storm surge which killed 300 people. So we were pretty worried about that and we put in place quite a lot of activity to make sure that that 300 deaths kind of scenario didn't happen again. Following on pretty much straight after that was a series of low-weather systems coming in from the Atlantic that we were just one after another and I think that there was nine or ten over that period that just came in and just chucked a load of rain on the country. Lots of wind, you remember the railway in Dawlish washing away, you remember scenes of flooding all over the country really. One of the reasons I've got notes is to remember the stats that I've got here now. So over that period we issued 50 severe flood warnings which is the top level flood warning that we issued. We don't do it very often because effectively when we say severe flood warning we're expecting people to immediately grab their stuff and ship out. So 50 is quite a lot. You can see it there at this point so the number at the top left there is two. That's two severe flood warnings at that point and you can kind of see the proportion. We're starting to alert people and there's a kind of like an order of magnitude shift. 8,000 properties flooded in that period, December through to May. 200,000 properties were protected by our flood defence assets. We closed the Thames Barrier more than 25 times in that period to manage the water flow in effect. Just to put that into context, the Thames Barrier opened in the 70s and in total over its entire life it's only shut 175 times operationally. To actually do flood risk management. So last year it was exceptional. However, the floods in 2007 were not over the same kind of period but were much, much worse in terms of impact. 13 people died and 40,000 properties were flooded in 2007. So why am I talking about this? This incident for us was not just a flooding incident, it became a political incident as well. The Somerset Moors and Levels story about flooding and dredging rivers in Somerset just became a massive political story. This led us to having what I'm calling our data incident. So alongside the flooding incident there was a data incident in the background. We got a phone call and it literally was a phone call at 4pm on a Friday. It happened to be Valentine's Day inviting us to open up our live flood warnings data which was a closed data product for a hack event on the Sunday. Bear in mind that the infrastructure we built to distribute our live flood warnings was completely and utterly behind our firewall. So we couldn't just flick a switch. We basically had to beg, borrow and steal resources to help us deliver that particular requirement. I've written down here a need to say there's lots of people internally that and we had a lot of help from a company called Shoot Hill in getting us through that particular period and that bit of the incident. Without it, we would never have got the data open by that Sunday. At that point, because it was Friday evening to Sunday, we said the data would be open for three months. However, it was unrealistic ever to think that after three months we'd shut it down again and lock it all back up again. Once the cat was out of the bag, that was it. So this fundamentally changed our approach. Although it didn't seem like it at the time, it changed our approach on how we thought about data and information. In as much as I mentioned contractual controls, one-to-one relationships to make sure that data is being used appropriately, if we can open up our live flood warning data, which let's be honest, if we get it wrong or people are using wrong information, there's a potential risk to life, if we can open that up, we can pretty much open anything up, can't we? So the only thing stopping us then is revenue and our structure. So we made a commitment in May at the end of this flooding event, which is coincidental, that we were going to become an open data organisation. So the first big change, now bear in mind that we've been doing open data for a while, the first big change was the release of this data product in December, which is what was known as the National Flood Risk Assessment Database. It's now known as the Risk of Flooding from Rivers of Sea Database. Why we changed the name, I'm still unclear, but we did. Between May and December, we were experiencing the long tail of the data incident. So flooding had stopped, we were obviously still recovering, doing lots and lots of work in terms of recovering from the flooding incident, but this was still going on. And the reason this was, we were still dealing with the request to change our data to open data as an incident, so effectively we were working in a tactical way. There were stacks and stacks of political interest in what we were doing, because obviously the flooding's there in the back of the mind when you're opening up your flood data. What are you going to do next? So we took a kind of project approach to dealing with it, which, on reflection, looking back, caused us more problems than it was worth, because what it did is it set a hair organisation that we would need lots of resources to become an open data organisation. We've got this data set here, how are we going to make that open data? How much is it going to cost us? How much do we have to invest? So that was all going on in the background, lots of politics, lots of internal conversations. Alongside which, we were working with Cabinet Office, Cabinet Office were putting pressure on us, or working with us, encouraging us, to release the National Flood Risk Assessment as open data. There's a lot of work involved in that, explaining to people in the Cabinet Office what it was they were asking for, working through what the implications were, and then working through how we are actually going to deliver it. It was December the 16th, I think, that we actually released this particular data set as open data, and I think we were surprised and relatively happy about the rapid proliferation of products that were built off the back of it, that we thought would happen. So us in the data world thought these products would develop. I think the rest of the business was kind of a little bit more unsure. This was quite a significant moment. This data set was a million pound a year data set to us. So flipping it over switched off a million pound of revenue. So it was quite a big deal. It showed that we could do it. It was quite a big moment. Alongside all of this going on, we were setting up a thing called the Environment Agency Data Advisory Group. So as part of our open data journey we were committed to getting the user voice heard. I'll come back to that in a bit. It was just before that release that we appointed Tom Smith as our independent chair, and I think our first meeting was just after, or just before, just after, around the same time. So that was the end of part one. Moving into the end of part two, so the start of part two, sorry, I think the release of NAFRA was effectively the end of that incident phase. We basically, given up that one million pounds worth of revenue, that reduced the political pressure on us quite a lot. We could be then seen to be playing ball with this open data agenda. It gave us the ability to stop and think for a bit, draw breath, work out, right? This is interesting. What are we going to do next? The two things that were important to me were making sure that we started to adaption because whilst open data wasn't a project and it wasn't going to be a massive additional burden, we did need to adapt how we did things to make sure that we were going to be able to do open data properly going forward. Alongside that, it was really important to maintain momentum. The pressure externally from Cabinet Office kept everybody very interested. It made the decision to make NAFRA open data very easy in as much as the pressure was there. As soon as the pressure comes off, it was, how do we maintain this? This is where the data advisory group, I think, started to play its first key role by having the user perspective engaged. These are people that are pretty reasonable and pretty balanced in the main, although they believe in open data. Ellen's laughing at me. On our site, we've got a group of people who represent our users. They're not just the people that we know, they represent a cross-section of people who are interested in advising us and helping steer us through this journey. My executive director, Miranda Kavanaugh, comes to that group and hears what the users are having to say about data, what we're doing, our approach and how we're moving forwards. Through that forum, I think steady pressure has been maintained on us to keep stepping forward every meeting that there is. Miranda's interested to make sure that we've actually moved forward and we've delivered what we've said we're going to do and we've got some good news to be sharing. So it's back to normal. We've got some time. We start planning. We start thinking about what's next. Working with the advisory group and our internal data owners, who we call custodians, for what reason I don't know, but that's the word that we use, we started prioritising across our entire data portfolio. So we've got 1700 defined data sets recorded for us in the environment agency. Some of those we know a lot more about than others. Some might be just a line on a spreadsheet saying it's this and we don't know what it is. There's about 300, 400 that we've got really tightly defined that we understand. We know the attribution. We know what we can do with it. We know where it is. Obviously that moving those lock, stock and barrel to open data and distributing, sharing them, putting all the infrastructure around them was quite a daunting task. So working with the advisory group and the custodians, we started to prioritise. So what was important to the users, but equally, what did the custodians internally think would be important to us in delivering our environmental objectives? So this is not just a journey about somebody's asking us for open data. We'd better do it. It's much more, how can this work for us? From that long list of 1700 and the short list of, I think it was a couple of 100, we ended up getting down to sort of like the top five. And with those top five, we can then start working through what are the obstacles? What are the barriers? What do we need to do? Who do we need to bring with us? Who do we need to convince in a much more targeted manner? Because obviously when you've got a big wardrobe data, it's really difficult to engage with all the right people all the time. Working with the advisory group here was really important because our thinking and discussions were done in public in effect. So the information was shared. It's been published on the advisory group blog and the addresses at the end of this presentation if you want to look at it. So by being transparent in that way, we showed the commitment that we were actually taking this seriously and we were moving forward. And again this gave us more time to breathe, to actually do it properly and put in place the right kind of process to get through. I've drawn a little picture on my notes here of an elephant because there's an elephant in the room through all of this, and I think it was one of the first things that was said when we started talking about prioritisation is we already know what the most important data is. It's the stuff that people pay for. Get that out open first. It's like, right, okay. That's a fair comment. So Charge 4 is our high impact data. It's our high value data because people have paid for it over decades now. So we know that that's the most important stuff. We had to come up with and we were still working on a credible plan to migrate our Charge 4 data to open data. The reason we need to plan, and it's not just like that, is to give up £5 million worth of revenue because bearing in mind, naffra's gone now so it's gone down by £1 million is quite a significant thing to do when everywhere else is driving austerity through. So there's not enough money in the public sector. Everybody else is, you know, we're all cutting, cutting back, retrenching and there's a £500 million, sorry, not £500 million, as if, £500 million pot of money that we're just going to go. No, thanks. Quite a bit of a journey to bring people on there. But we've got there. So a month and a third of June this year, I took a paper to our executive directors team. So that's the top of the organisation in effect. That group's chaired by our chief executive and that paper had two key bits to it, which is it outlined a transition plan and got them to agree to a transition plan that all of our charge for data would be open data by April 2018 with two steps in between. So effectively a third of the money would be given up next April, a third of the money would be given up April 17 and everything else goes April 18 and after that we don't charge for data anymore. The reason for doing it like that was it was just too much to swallow in one hit to give up the £5 million. But this picture illustrates the other part of the decision that was made that day, which was that all of our LiDAR data, which I think it referred to say the feedback has been probably our most important data set in terms of its multiple uses by multiple people, will be made open on 1 September this year. It's just worth reflecting on the discussion that happened in that room about that decision. So there it was on the piece of paper that I'd written with some help from other people. I'll claim a credit. Outlining how we will be giving up a significant revenue stream from this very, very valuable data set and we'll just give it away. And the start of the discussion was pretty much about how much does it cost us to collect all this stuff? Why on earth would we give this to people like the Ordnance Survey who are then going to make money off the back of it? Why would we do that? It's mad. But then I introduced the conversation or steered the conversation towards what would the impact of making that data open be on that area operational staff? So these are the guys that sit in the area all around the country. If a developer wants to build right next to a river, they have to do a flood risk assessment. They send that flood risk assessment into the area guide. They then need to go through it and check whether the model that they've used is correct. We agree with it, whether the proposals that they've made and mitigate flood are OK. And then they approve or say no chance. At the moment, the data that people use to do flood modelling is not just from the environment agency. There are a multitude of people that do flood modelling in this country. So they could be using data from any source. I can pretty much guarantee you that they won't be using this data to produce those flood models. So the argument that I put to the executive team is by making this data free, those flood models will be improved by using the best quality height data rather than stuff which is at one metre or two metre resolution. Some of this is 25 centimetre resolution. That makes the people in the area's job easier. They know that the models are built on the best data. They will then, it will be easier to trust them and therefore it will save us a bit of time and a bit of effort. So by making the point that this is not a completely altruistic gesture, we're not just giving it up for the good of the world, but actually it has a direct, significant environment agency benefit. They just went, well, we'll do it then. It was pretty much a turn round like that. On the basis that we would get benefit, even if it wasn't monitoring. What that decision a month ago gave was me and my team a mandate to start implementing some of the changes that we need and putting in place the plans. So, on to stage three. And this is the bit that I've had to work through over the last month or so, in particular over the last week, just to make sure it tallies with what's been going on. We need to change the culture of our organisation. At the moment, if we're building data or we're coming up with the idea for a new product, we work very much in a silo. We are the user community for that data. We build it for us. We might be building something for somebody else, but then we will only speak to that somebody else. So, for example, the insurers. We might speak to the insurers about what they want, but we pretty much focus on what we think. We need to start changing how people see themselves, so we start to see ourselves as part of the user community. What this is, is part of that kind of first step. This is our new data strategy in the environment agency. And when I say this is it, this is it, seven words. So, this is a distillation down of a 15-page document, which then became a five-page document, which is now seven words. In this room, you'll read these four words across the bottom, and they'll mean slightly different things to everybody. If I set you what does our data is valued mean to you, you'll give me an answer. Well, that means that people are using it to make the right kind of decisions or whatever. What I can guarantee is all of the answers will be a little bit different, but they'll all be pretty much on the money of what I think to, you know, within the broad gambit of what data is valued is all about. So, we want people to actually feel what the data strategy is rather than tell them what the words mean and tell them what it means to them. So actually, by having these words interpreted in your own mind, it's much more likely to stick, you're much more likely to believe it, and then you're much more likely to act on the back of it. So, that's your own interpretation. I don't tell you what it is. Hanging off this is our governance framework. That's where we start to use things like the environment agency's maturity model. We started recently just done the Open Data Institute maturity model. Using both of those as tools, we're starting, we've obviously benchmarked ourselves, and we're now starting to refine our approaches and process to data management, both on the basic data management side, and also sort of preparing for open data within all of that. The idea is to move from this kind of siloed approach to be much more holistic in how we think about data, the users, and the final use of it. An example here is currently what happens if somebody is creating a dataset, they'll go away, they'll set up a project, they'll run that project, they'll build up a load of data, they'll get a load of inputs, they'll build something, and at the end of it, they'll say, right, we've already satisfied that internal use case, hurrah. And then we'll go up to them and say, well, so and so has asked for that. Can we share it? And we'll look back through all of the processes that they followed and the licenses for the data that's come in, and find, oh, no, no, you can't. We do that check at the end. Part of this culture change is we need to build a process, which I've nicked from a person called Lisa Allen, this term, our prep prepare to share process. And that is working with people that are building data right at the start to make sure that all the way through, all the decisions that they're making, they're thinking, can I share this thing at the end? My little advert is Lisa and I have written a blog which is going to go on the ODI website on Monday, hopefully. Well, I said I don't go on Monday. So there's more information on how we're going to use the maturity models to actually shape our work on that blog. So have a look at that. It's not just about an internal change. Back to the data advisory group. We need to utilise the data advisory group, not just to tell us what they think we should be doing, not just to prioritise our data, but to actually be the stakeholder group that we've never had. So quite often we'd come up with a project, and part of that project plan is you need to engage with stakeholders to find out what they want. So we'll go and speak to the people who we know what we're using the data or we've worked with before, and they will tell us exactly what we want to hear. Yeah, we need a new map. Go and build us a new map, hurrah! Well, actually, that's not what's needed at all. So we need to find a way to combine our business needs with the user needs, combine our business users with the wider user community to actually understand what people as well as us want to see if that can be built into what we do. Fundamentally, the most important users, frankly, are us because we're collecting, creating, using data to do our job, but we can consider this group. There's no reason why we can't, and the data advisory group and other groups that we already network with can actually provide that input usefully. An example of where this has worked already is during that flood event that I mentioned last year, you may have seen, it was on Sky News, on BBC News, on our website, the live flood map that Shoot Hill had built. We got a copy of that on our website, and we were looking to replace it, and the reason we were looking to replace it is because it was hooked in through our firewall and we wanted to build something that we were in control of. But we were looking to replace it, and we started off, well, we need a new flood map on our website, don't we? But we did, we followed the GDS user story approach, went out properly, consulted proper users rather than the people that we knew, and basically the feedback was, we don't care about a map, we just want to know if where I am now is going to flood. That's all I want to know. So if you go on to the website on gov.uk now, which relates to live flood warning, the first bit you see is do a search, find out if you're at risk of flooding. You can get to a map, but the first bit is that engagement. There's a link to that at the end as well, so you can have a look at that. The next bit is about trusting these people. We've been grown up as a teleorganisation, populated with experts, with data experts. We know what we're talking about. We do rainfall, we do weather, we do hydrometry, we do all that stuff. So we can tell you what you need. The change that we need to sort of move through is actually over the last 10 years, there's a whole group as shown here of people who know how to do data as well. It's not just us anymore. We need to trust people that they can, if we just give them the raw data that we hold, that they can do something with it, they will understand what it's suitable for, if they read the materials we provide with it and they can be trusted to do sensible stuff. Any poor use will be spotted, outed and discredited, rather than policed by us. So, what else? This was me on a cooler day, and I don't mean I was cooler, I mean the temperature. Doing more of this, talking to people about what is the journey that we've been on. It's actually providing the resource of our experience to others. So across government, people are at different stages embarking on or well on this journey. So I've been to talk to, this was the inside government seminar, back it, I think it was at March, did in March. I've spoken to OFCOM, just done stuff with DEFRA, Centre of Ecology and Hydrology, British Geological Survey, Ordnance Survey, about what have we gone through on our open data journey, what's important. And it's not just the stuff that you would imagine. Had a really interesting discussion with OFCOM about how you work with lawyers. It was a fascinating discussion. So, more of that. And kind of to conclude really, the future is, open is just how we are. So we've got this plan, we've got to follow the plan through and we've got to get there. And then, hopefully, where our Secretary of State gave a speech talking about DEFRA opening up 8,000 data sets. We're now part of the project that he's looking at, how we're going to do that. Sharing our experience into that project. It's an ambitious target, but I think it's an achievable target because the will of the top is there. And that's the critical bit. Up until now it's been very much the data geeks, no offence intended of course. Data geeks, very much believing in this and trying to push the agenda. Now we've got somebody who's basically the top of the whole show telling us that this is what we're going to do. We're going to be sticking with our current plan because DEFRA have said they're happy with our current plan, so that is moving our data open as quickly as possible and the charge for data by April 2018 all being open. So that's the plan and we're going to be sticking with that, but this is where we're going. The kind of final bit, so what's the next, when I do this presentation in another year what's the next bit of the journey and I think that's going to be the most interesting bit because that's working out who, how and when we start to work in partnership with other people. So rather than thinking we need to build things ourselves or think that we're the only people that know how to do stuff, when the resource, our data is open people already be getting ideas and thinking about how to develop stuff which will actually help us. How we pick and work with different people is going to be the next big challenge because there's going to be lots of stuff out there. How to use open data to better deliver our organisation outcomes. And that's kind of where I want to end.