 Good afternoon everyone so I'm just going to give you a bit of an overview of the energy industry and some of the things that are happening there and then a examples of different use cases that ESB has implemented to be a bit smarter and user data better. So just to introduce myself so I went to Trinity and did an undergraduate I was quite interested in data so I went on to UCD did an MSc and then I went into Care Group and then Bank of Ireland and then finally ESB and I just a year. So you can see there are three quite different industries but they all seem to have the same problem of using their data to make decisions better. So surprisingly ESB was actually quite a day-driven company. This picture is taken from ESB archives and so it's a pencil and paper diagram of the flow of the river Chatham and it was taken when they were developing Arduino Crusher and back in the 1930s so about 90 years ago and thankfully they did move on to digital records and using computers and probably sooner than the 1990s and now today data is kind of quite key to a lot of things we do and we've also moved on to looking at unstructured data as well as text to try and understand sentiment and also data is now included in a lot of the products that we use so our customers in Electric Ireland get meter readings and trends in their monthly usage and over time so the energy industry itself has undergone quite a big transform so the traditional energy industry energy 1.0 would have been where just power was flowing one direction and also information was flowing one direction we had quite traditional generators like coal oil and gas the meter in person's house was just a clock that was read occasionally and also the usage of energy was for quite simple things like kettles like bulbs and probably ovens and but now energy industry has transformed it's going through quite a big disruption where everything's a lot more digital and the IoT devices are playing a huge role in the homes so people are now actually able to see their energy usage for each device they're also able to monitor their heat pumps and reduce or increase their heat to reduce or increase the energy they're consuming houses can also now produce their own energy which I don't think was something our grid initially thought would happen and but now it's being rolled out at such a big level and volume that the network side of ASB is looking at how do we make our grid stable so that when 50 people plug in their electric cars that the lights don't go out in the whole estate and and that's finally the energy industry so through ASB so far we have managed to work with a lot of different parts of ASB because it's vertically integrated so we've managed to work with electric garland a little bit as well as with generations and networks we've also worked with a few of the startups based down in planet line and being one of them and home hero and certainly plus and we've also started to look internationally at with ASB international at the international blinds and also you notice PA the house is up there so as part of our corporate social responsibility and we did a project with PA the house and to try to help them to make more data-driven decisions and to get more benefit out of their data so this is just an idea of the different value levels that you can gain through your organization and so the first value is looking at the data and trying to understand what happened so that's just describing an event second is using the business understanding to try and diagnose why did it happen does this actually make sense what the data is saying and this is usually done by bi reporting team but sometimes there might be a bi reporting team so for a project we'll bring a client the whole way through from looking what happened to why it happened and then maybe into predictive or machine learning and as we would like to be doing and so these are different stages I'm just going to give a different use case for the four stages and and then just within the project we've worked on so far there's kind of two the top two are mostly based around resources and like asset optimization and the predictive maintenance that's mentioned a few times today and and also we look at safety and behavior this is again different types of data and also then in electric garland we've looked at customer insights and segmentation and so this is an example of the first level of a descriptive project so all the data is being stored on a share point site and this stored I think was bids that ESB had made for international clients and there's about 120 countries involved I can't remember how many clients and all being stored on a share point it's quite difficult for the bid manager to get a handle on how many bids they had and what stage of process they were in and so what we did was we worked with the domain expert who was our bid manager who was luckily also our end user and to build a dashboard that would help him to actually see big picture and so before we built this dashboard and he spent about three to four days a month compiling all the bids into an Excel and then creating a presentation so we've given back few days each month and and what he can see now quite quickly and easily and in real times well is where bids are at and actually when a bid is due so he can see that a bid submission is coming up or that he has five bid submissions coming up and is going to be really busy and also the other benefit that he's realized is he can see in a specific region like Asia that he's not winning a lot of contracts for a specific product but he's actually winning a lot say in the Americas so he can then change his focus or change his teams in the areas to reflect that so the second type project that we've worked on is trying to use the business understanding to validate what we find in the data so this was trying to understand our competitors cash flows of revenue for a specific product called DS3 products and so what we used was publicly available demo data and that's based around generations and availability so basically if you have a difference between your availability and your generation you get paid for one of the products and so this was used in Python to develop the code behind it and then we just presented it in tableau for the business users and this probably wouldn't have been possible to do without the business because the rules were so specific to calculating the revenues so that what this let them see was to gauge how much revenue their competitors were making and to see their different strategies and maybe where we weren't doing as well as them and could probably change our strategy for the better and the third project that we did is Planet 9 so this is a startup they're just entering into the UK market and what they're going to do is they're going to offer their customers the ability to buy energy at wholesale prices so for a customer that sounds great because the wholesale market is much cheaper than the commercial market that they will be buying off traditionally but for a customer they need to know how much they need to buy so they can lock in their price or they can hedge it and so for this we use smart meter data and one year smart meter data per customer to try to forecast their usage of energy and to let the customer make the better more informed decision and so this the ultimate benefit of this was to reduce the cost for the customers and to keep them happier this is third or fourth use case that I have and so this is a descriptive one that's still ongoing in development and so we're working with ESB networks to ensure that we can try predict when an outage will occur and the severity and location of the outage and so what we have is we have weather data and also historical outage data the outage data includes reasons for outages just weather or fault so we're just looking at the weather ones and where they occurred to try to identify can we get down to say we need five crews in cork and because the weather is indicating that's going to be the worst hit as opposed to having the whole country having five crews because we're not too sure where it's going to need it and so again the benefit of this would be minimum disruption for our customers and restoring service faster and again a better customer service and so then finally how we organize our projects so now we use a project cycle and generally the first step is we'll export the data then we'll go back to the business to present our findings and ask them for a bit of guidance on is there anything else they'd like to see and then they'll have some further questions or specific areas and then finally we do a handover and review session so for us it's more important to do an initial proof of concept and explore and present back to see if we're going the right direction or if it's useful and then after that feedback to keep iterating through it and so that's really all thanks