 So before I get started, I would encourage you to visit the DATSY stand. They have a stand outside Linda and Emma and the team get involved in the DATSYs. It's great that we have a community now that we can actually share the projects, what we're doing and go up against our peers. It's the third year of the DATSYs this year and highly recommend people get involved in it. So I'm going to talk about something less exciting than the DATSYs. I give my own perspective on the application of lean technologies applied to data science in supply chain. Manufacturing in Ireland is a big deal. It's over a third of our GDP, which the European average is 15%. We employ 10% of the workforce directly in manufacturing. 20% when you take in indirect manufacturing of the workforce of Ireland is employed in manufacturing. 70% of those are in indigenous companies, only 30% in multinationals. We spend over 771 million every year on R&D. Over 8 billion on materials and services. And we contribute over one third of the country's gross added value way above our European counterparts. So manufacturing is a big deal in this country. But manufacturing is changing. Digitalization and digitalization is changing the face of manufacturing. And if we don't get on board, we will suffer because of it. The World Economic Forum states that over the next 10 years, over $100 trillion will be generated through social and industrial impacts. But they also estimate that anywhere from 2 million to 2 billion jobs can be lost in the next 12 years due to automation, digitalization, and a lot of technologies that we've heard here today. So digitalization, whether we like it or not, is going to have a massive impact on our manufacturing and our wider supply chains over the next 15 years. The European Union estimates that $110 billion of revenue will be generated in the next two and a half years out of digitalization. So digitalization is a big deal. And it's going to impact on our supply chains and our supply chains have to adapt. They're going to have to do things differently. The same way as when we look at companies like Google and Facebook that have their European headquarters here in Dublin, they're offering jobs to graduates that didn't exist 10 years ago. We're going to have to do the same in supply chain. Kids finishing secondary school now and students coming out of college would be looking for jobs in five years' time in the supply chain that we don't have today because we didn't need to have them, but we will need to have them in five years' time or less due to the impact of digitalization and analytics. So traditionally in Ireland, when we think of manufacturing, we think of a single node in the supply chain, but you've got materials coming from somewhere, you've got to deliver it somewhere, you've got all that returns policies, and you've got a plan to have that, and that's just you. And the amount of data that you're creating is massive, but you're just one node in the supply chain. So every other player in the supply chain has got a similar supply chain to you, and these things just get out of hand very quickly and it's like unraveling a piece of string. They're incredibly complex organizations, and I would contend that supply chains are so complex that they are a rich breeding ground for the application of data science because there are so many problems that need to be solved and now that we're moving from brownfield sites into digitalization, we now have data on tap to go and solve those problems. But I would also say that data is tricky in supply chains. We have lots of data. We confuse collecting data with doing big data. Data can be good and data can be a waste of time. Are you collecting the right data and what quality is it? If anyone saw Martin Perry from Microsoft this morning, he talked about data fluency, and he talked about what's the question. I often hear about we did a predictive maintenance project and we collect with 26 sensors on 10 magic grinders, and we're going to collect terabytes of data. That's great. Why does 26? What question are you trying to answer? Maybe you are collecting the right data, but just collecting data for data sake is a fool's earned. You may collect it all over time. But what we also see in supply chains is data starts in an ERP system or an MES system, but it's then twisted and turned into pivot tables and hundreds of Excel files that are now seen as data sources, but they're not. So we have lots of data in supply chains and sometimes it's about winding it back and say, where did that data originate from? Because there's only a limited number of data sources coming out of a supply chain. So now you want your metrics to stabilize the data flow and create metrics is a big job. It's only money. It's only technology. It's not that difficult, but is that what you really want? Do you really want me to go and put a team in place to track these five metrics? Are these even the five metrics that you want to know? Because I've got to put a full operation there to stabilize that flow of data. What are you going to do with it? I'll give you some insights. What are you going to do with the insights that we have gotten from the project we did to collect data and stabilize that flow of data and make it flow? And maybe I can dip in and out of that flow of data now and I can do some maybe anomaly detection or some advanced analytics, but right now we're still only talking about descriptive diagnostic analytics here. But what insights are you going to derive from that? And then more importantly, and this was talked about again, this is a current, a common theme I'm seeing today, with that insight, what action are you going to take? Because if you're not going to take any action, why did you look for the insights and why bother collecting the data? That's a waste of resources and a waste of time and why are you doing it? And it all comes back to what's the question you're trying to ask? We see commonly in manufacturing the onset of digitalization is scaring people because the wall is getting bigger and bigger and all they're hearing about is this sexy AI stuff and machine learning and neural nets and they think, oh, we're so far off that and they are so far off that because they have to look at the basics, the digitalization, getting the data, stabilizing it. What happened in my manufacturing process yesterday? Talked about control towers earlier on. Do I have end-to-end visibility of what's happening with my suppliers? What my plan is telling me to do? How did I manufacture? Did I manufacture what my customer wants to order? Am I an agile supply chain? Am I a responsive supply chain? Do not have those types of supply chains. This is one that really gets on my wake, I have to say. A company goes into a VP in the US and says, I've got a big red button. It's going to do everything for you. It's AI in a box, you plug it in, it's going to do the Divlin Alt for you, 10 million bucks. VP goes, I'll have two of them because, you know, I just don't have the time, it's very complex. Give me two of those buttons and it's going to fix all our problems, third-party solution, plug it in, plug it in and away you go. That sounds brilliant. Go into my manufacturing site there and plug it in. Then they rock up at the front door and this is what they see. Where's your data? There's no electricity in this place. There's no data flow. Hang on a second now, we're going to need to bring in the SMEs and we're going to need to bring in half of your company just to stabilize the flow of data so we can plug in the big red button. Somewhere in the middle, it's just about having an honest conversation and saying, we want the big red button, but it's going to be a journey before we can plug it into the wall and press the big red button and that's okay. That's perfectly fine. It is a journey. Analytics in supply chains is a journey and you've got to get on board sometime because digitalization is steam-rolling ahead and if you're not on board, it was mentioned earlier on, you will perish. You will not survive the next wave of the technical revolution. So one of the concepts we talked about, we are a shingo in Depew Sinti's in Cork, we are one of the shingo award-winning companies. You get a shingo award for lean excellence. So trying to sell analytics in a supply chain is a tricky one because culturally it's like now we don't have to worry about that, but now all of a sudden we do have to worry about it. So we say, okay, these are your lean processes. You've got all these beautiful lean processes and that's great. But the data flowing through your lean processes is anything but lean and if you're filling data lakes, you're filling data sewers because you have data everywhere. You have hundreds and thousands of files. You haven't ever looked at the data flowing through your processes. You've just looked at the process. So if we take the concept of lean and I'm not a lean expert and I wouldn't disparage anyone in the lean community by trying to say I am, but if you take some of the concepts of lean and applying to the underlying data flowing through your processes, you could be much more effective. Lean and analytics. This was one thing we found. How do you adopt analytics into an organization that's culturally resistant to it? Well, most of those organizations have process centers of excellence. They went on a mad blast a couple of years ago sending everyone off to do lean six sigma courses, masters in lean process, all of that. So they get lean. They're probably leaned out at this stage of wanting to invest money in some other courses. So if you think there's a lot of similarities between the rollout of analytics in a supply chain and the rollout of lean in a supply chain. Lean is a toolkit. It's a set of tools and techniques to improve your business process. Data science is also a set of tools and techniques. It's like the plumber calling to the house. You ask him to fix the sink and say, by the way, the shower is broken. He goes, I'm sorry, I can't fix showers. I can only fix the sink. And it's the application of that toolkit across data flows in your supply chain. And even the journey on data science. So you might start off, this is your typical journey on a lean process improvement. So maybe you say, Joe, we want to move from Excel to click or Tableau or Power BI or some other BI tool. So we're going to upskill on BI tools. And then our green belts are going to become citizen data scientists. So these are not data scientists. These are people that work in the business sector, but want to upskill on some of the data science tools, maybe Python or some ETL tool or whatever it is to get more out of their data. Then your black belt might be your data scientist and your master black belt might be someone who specializes in machine learning or deep neural networks. But you don't need too many of those in an organization. Or maybe you don't need them at all. So there's a lot of similarities with how you would embed data science and analytics into a supply chain if you follow the lean trajectory. So this is a very simple app, a suite of apps that we created for one manufacturing site in Depew Cinti's. And again, it was using that concept of applying lean tools and techniques to the data working with process owners to say, what are your biggest pain points? And instead of building one big over arching software application that's out of date, what we would do is we would build it very small. And these small releases, small types of apps and embed them in the business because they were being led by the process owner. So the suite of apps would consist of capacity planning optimization tool and MES, the work orders that are flowing through my manufacturing process. Where are they at? Are they overdue? What stage are they at? And the scheduling application. The key thing here was we didn't integrate into any systems because when you're in a validated environment this can kill off a project saying it's not a problem. So all of the files that they used today in our manufacturing processes, they were our data source. Whatever files you use today, we will map to them. Then each one of those apps had a different modeling component, so it might be a mixed integer programming component for optimizing your capacity plan. It might have been a probability rule-based tree because you had a non-linear manufacturing process. And at this point you were able to show put all your data then into a file and send it out to your BI tool. And this was all very seamless. And they controlled it because what they would do is they go onto the website, download the files that they were already using today, put them into a folder, press a button and wait for it to process. So they controlled the flow of it. If the system went down, the system was down, and I don't mean this system, but if their MES system was down, they couldn't generate a flow of information down because they controlled the flow of information through the system because everything was on through the files they were already using today. And then if they're getting benefit out of that, then you can up-rev it and take it on to integration into validated systems because that's a different animal altogether and we want to stay away from that. Benefits, reduction in processing hours, processing loads and loads of Excel files, but the real benefits is what work orders are overdue. What work orders do I predict will go overdue in the next two days? What should I be working on? Because it was mentioned again today, what's managed, what's measured gets managed. And then capacity planning, being able to have a conversation across the supply chain, a data driven conversation to say you can't take that machine offline for two days because if you take that machine offline, that means I have to run my manufacturing process and the customers are currently looking for. So it becomes a much more fact-based discussion. And finally, with 10 seconds to go, I will leave you with a famous line from a famous management consultant culture is a strategy for breakfast. All of the analytics in the world isn't going to help if your organization isn't ready to adapt to digitalization, doesn't see it coming down the line and doesn't have the right culture in place. is a maturity mindset to say, this is happening, how am I going to embrace it? I might not be ready for deep learning or neural networks or any of that sexy stuff, but I do need to know what's happening in my manufacturing process, and I do need to see in my control tower what's happening across my supply chain. Thank you very much.