 Thank you. Thank you, everyone. It's great to hear such a buzz in the room and I'm sure that that will continue throughout the evening. And good evening, everyone. My name is Jane Godfrey. I'm Dean of the University of Auckland Business School and it's my absolute pleasure to welcome you here this evening for our short, sharp event. And just as a reminder, this event is brought to you, of course, by the University of Auckland Business School Alumni Relations Team and also our Graduate School of Management in Collaboration. And for your information, I think some of you already know this because there are a few paces here that I've seen before. In fact, some that are quite familiar to me. And it's great to see you here because we at the Business School really do believe that it's important to maintain our linkages with alumni. And we want to connect you with the Business School but more so with each other as well. And when I talk about alumni, I use the expression in the broadest sense. So yes, alumni who have qualified with formal qualifications through the university, through the business school, but also those of you with whom we have a relationship for other reasons. We really appreciate that it's important that we maintain our contact with you and help you along your journeys. And so that we achieve our own missions. And we do appreciate that in the current world of global opportunities, you have plenty of other opportunities. You could be somewhere else tonight, but of course you've chosen to be here and you have chosen wisely, I am sure. So you could be anywhere though, day or night, and you can actually connect with us by social media and other means. We see it as our responsibility to provide you with meaningful opportunities that will inspire you to want to maintain your contact with us to help us achieve our objectives. And events like this are a collaboration between alumni relations and between executive education, which is an arm of our Graduate School of Management, because we recognize the importance of professional development. And we recognize its importance to you as professionals, to business, to the future of management capability more generally, and of course to the New Zealand economy. Whichever category of alumni you fit into, I am glad that you're here this evening and I hope that you'll take away something new that will help you in your professional lives. And I do, as I said, encourage you to stay in touch with the school, with the business school that is, to make the most of our networks and events like this. It's now my very great pleasure to introduce Devon Dean, who is of course your speaker this evening. Devon Dean is an Executive Education Facilitator. He is the Director of Data and Analytics at Enterprise EIT. And in this role, he basically ensures that customers can easily navigate the complexities of implementing data warehouses, data management, business intelligence, and business analytic tools and solutions. His experience ranges from managing multi-million dollar projects to leading teams and starting businesses. And he attributes his success to being proactive and to having an inclusive management style. Devon's 25 year career spans such roles as being in the United States Marine Corps, working with tech startups in San Jose and Orlando, and locally the Simple Group and Business Intelligence Consulting firm Altus. And outside of EIT, Devon is active in New Zealand's startup community, the tech startup community. So please, I think we have a very well qualified speaker tonight, please join me in welcoming Devon Dean. So welcome Devon. Thank you, Jane. Appreciate it very much. Right, before I get started, I will give you some warning about how I present and how I like to teach. I'm very interactive. I'm pretty full of energy. So if you came here just to watch, take some notes, you're in the wrong place. I can tell you're going to be very active tonight. So hopefully that's okay with everyone. Also, a quick show of hands. We started a little bit late. Is it okay if we go to about, you know, 735-ish, move past the 730 mark? That's okay for people. Okay, great. And a quick show of hands as well. Those with a technical background. Raise your hand. Okay, so about a third. Awesome. Right. So keep me on my toes and make sure I'm not telling any lies, if you will. I often speak to crowds ranging from business to technical. I have to say thank you very much to the University of Auckland. This is a great reception to come into. The last presentation I gave was a couple of weeks ago down in Wellington and I had to fight through a crowd of protesters at the Defence Industry Association Forum. So this has been a nice, pleasurable experience so far. Okay, so to get into it, I'm going to teach a little bit and hopefully you're going to learn some knowledge nuggets from me. And I'd love to learn a little bit from you as well. Can I have your permission to learn from you? Yeah? Shaking heads, yes? Anyone who's not okay with me to learn anything about yourselves? No? Okay. Cool. So the first thing I'd like to do, many people when you go to presentations or workshops, they say look, please put your phones in your pocket. I don't want you to see them, whatever. That's not here. I want you to take your phones out if you can. Yeah? Go on. If you have your smart phones going, those who have them, it's completely optional. But those who have your smart phones, pick them up and I'd like you to go to a website for me first. Okay? And the website name, it's really easy to type in. It's demo, d-e-m-o dot e-i-t. That's where I work, dot n-z. And what that should do is take you to a Slido area. Are you all there? The website address is demo dot e-i-t dot n-z. Yeah? Should take you to a splash page, then you have a link to go to a Slido. Did you get it? Yeah, all right. Cool. So I'm really interested to know why you sat, why you sat. That should be the questions coming up on the Slido. My colleague and two IC in the back, John, would have enabled the Slido poll going. Is that working? Yeah, cool. So did you sit there because you have friends? Did you sit there because you wanted to meet new people? So if I could have one person from each table stand up, please. One person from each table stand up. Come on. Three, two, one. Let's go. Quick, quick, quick, quick. One person each table. Great. Volunteers fantastic. Oh, is the poll working now, is it? No. Okay. We'll do this first and then we'll get to the poll. Everyone else at your table, can you please take three sheets of paper from in front of you and sort of scrunch them up into balls? All right. And it's important that you get three. The person standing up doesn't have to do this, but everyone else scrunch up the paper into balls. Three balls. Yep. Every person gets three balls. Hold on to them. Don't throw them away. No, person standing up doesn't need it. Okay. You, everyone have three balls? Every person has three balls. Not the table having three balls together. If you've got a table of four, you should have 12 paper balls. A table of five, 15. It's easy math. Yeah? There? Great. Okay. So before we moved into the age of the internet, the way that we interacted with our banks, insurance companies, and the way that they interacted with us was a little bit of a slower pace. So what I'd like to do is the person standing up, you're gonna have to test your hand-eye coordination. And going around the table from left to right, each person will throw a ball, one ball at the person and they'll toss it back. So go on. One ball to the person and toss it back. Yep. Nice, easy pace. Easy to deal with. Awesome. Awesome. Okay. Stop throwing the balls around. That's good. That's good. So what's happened? That's how we used to work with our, thank you, cheeky. That's how we used to interact with our banks, insurance providers, hospitals, however. In the last 10 to 15 years, we've had a lot of different things have changed. We've had the internet roll out. We've had lots of different systems such as CRM systems, social media come at us, right? And now, not only are we expecting to interact at a faster pace with our institutions at service, but they're also trying to interact with us to match that pace, to learn more about us and provide more service to us. So at this time, I'm now gonna show you how things work now. The person standing up, right there, everyone all at once throw all three balls at them and they'll try and throw it back, one, three, two, one, go. Yeah. Right. Okay. Good. Have you got it? This is hot spotted. Okay. Thank you. That was great. You can thank, round of applause for the people standing up. Yep. Great. So with that visual metaphor, hopefully you can get an appreciation for your service providers, the banks that are working for you and also yourself. Remember that you're having that same demand on them. So it's really, really daunting to try and build up a relationship and actually not only get your service's needs met, but the institution you're working with, having them understand you to make, for them to know what services you're asking for and how better to meet those needs. Have we got the poll going? We still don't have the internet access. Okay. We'll move on and we'll get back to that one. So why be data driven? So you can understand with that visual metaphor, the balls being data going back and forth, that it's really important to be able to handle that data, right? So right now, if you could open up your workbooks on the first page, under the topics, everyone have a workbook? Cool. Just off the top of your head, list four, five, six, whatever you come up with in terms of why do you think we've got a problem today with data and working with the service providers that we're trying to interact with and for them to build up a relationship. Do we have a problem at all? Or what are the opportunities for organizations today? From what you know in your own work life, what would you say those problems or opportunities are? So go ahead and write those down. So after you've finished listing out one or two topics, please go back to the demo.eit.nz site, and then that should take you to the poll you put on the boards. Cool. And look, be honest, if you're trying to avoid somebody or trying to get out of the get out of Derek, the presentation quickly, then please be honest. So to let you know, there's no information other than your answers being collected, right? So we're not linking this into Facebook yet, or LinkedIn. Cool. Right. So good view, right. So what have we got polling? Most people say that there was a good view from the, yeah, some people are meeting new people. Great. Speedy exit out. I love your honesty. That's good. No one opting for the I'm avoiding people option. Liars. You're all liars. Okay, fantastic. Okay, that's good. So today, what we're going to talk about is about why be data driven. And what I tried to do in the first five to 10 minutes here is really set the scene about, you know, in a visual metaphor, what is the problem that organizations are facing today with all this data coming at them from multiple places. I'm also going to help you understand what to look for in a project or an initiative that actually might help you run your business a bit better or interact with your customers better using analytics. Of course, the foundation of those is data that pull that gets pushed into those analytics. And lastly, we're going to help you select some of those ideas that you might have and choose an analytic project. Now today's short and sharp is a very condensed version of a full day workshop that I run with the university. And if you want to learn more or come back, come on that training to get into the details, please join us then. But hopefully after today, you'll have enough knowledge to go back to your organizations and maybe start triggering some of these ideas and putting them into motion. That's my hope today as well as a little bit of entertainment. Hopefully, most of the kinks have been worked out so far today. So why be data driven? We talked about that. And now I just want to set the scene in terms of what I guess a general definition widely appealing definition of what analytics is. And effectively, when people talk about analytics, they're talking about finding ways to understand the data, looking for those patterns, right? As an example, in the previous days, and anyone work in the marketing crowd, anyone here doing marketing at all? Yep. So you're well aware of segmentation analysis, understanding the customers a bit better. Well, generally in the old days, that segmentation analysis was as good as we could get because of the level of data that we had available, or the ability to process that data limited us to be able to create categories for these datas. Nowadays, with the processing capability we have in things like big data, or even in the cloud, you're able to process a hell of a lot more data. And as a result, the clustering technique is being overcome by, excuse me, the segmentation techniques being overcome by what's called clustering. So ultimately, in a perfect world, you might have a cluster of one, an individual. But generally you find that patterns of behaviors exist between people, and those clusters can come together when you're using these large big data solutions and with the modern analytics. So that's what we'll talk about today, hopefully to tune your radar in to pick up ideas from your own organizations you can take back with you. So the little bit about the chronology. And unfortunately, I don't have a great story to tell about the ability to actually execute on those ideas that you might be having. In the past, as people started gathering data, it was good enough to see what happened yesterday. And I got to tell you that's really still important. In the industry today, I hear a lot of vendors talking about, forget about the yesterday's reporting, who needs a data warehouse or BI system, you should be doing predictive stuff. If you don't actually know your history, how do you know where you're going, right? If you try and overwrite that history, you can't actually understand that you were at this point and now you're here in the redirect. You can just be caught in sort of a circle of confusion. So I have to tell you that those techniques and tools and way of working is actually still really important. But what's more important is learning from that history and then moving that forward. So over the years, as technology increased and new methods came available, you could see that it became a lot easier as a business person to actually do this myself. And fortunately, it's turned around a little bit. So it's very codey out there right now in terms of software development. If you really want to tap into this machine learning or what they call advanced intelligence or artificial intelligence, you still need a software developer to do that. The tools haven't matured enough to keep pace with that. We did have a small turn when the data visualization sort of revolution took over. People may have used tools like Tableau or Click or even Power BI. And then it became a lot easier to actually work with the data yourself. But nowadays, I have to tell you, it's kind of gone backwards a few steps. Probably because everyone once has higher expectations of what they want to achieve. And also because of the revolution of how these things are being developed in an open source community rather than being pushed by big vendors, which is actually really democratic. And you see a lot more technology evolutions happening. Right. So let me just give you a few categories under this larger category of analytics, right? I break these analytics into four different areas, right? The first one is business intelligence. We talked about that earlier. And really that's about, it is about looking backwards. You can't do away with it. You absolutely need to know that history. It's especially important when you're doing regulatory reporting. Finance banks, they have that requirement. Insurance companies have to keep data for seven years, for example. You know, that is very important and will never go away. That's really your dance card, to be honest. Because in a maturity matrix when you're looking at, hey, where do I go for my organization? If you try to start doing machine learning right now and you don't have those necessary ingredients to get there, you're going to stumble and fall. So you really got to start at the top and sort of work your way down. So doing some business intelligence is really important. The next area of analytics is data mining. In that you're actually exploring. You may not know what you're looking for, but you've got the algorithms at your hand and you've got the data sets to interrogate that actually show you what those patterns and behavior behaviors are. And you might, and I've given you some examples here. So, you know, BI is like, well, how many customers are buying this product this month, right? The next one in data mining is, well, who's actually buying the product? What do we know about the purchaser and, you know, what are they, what are they buying it and why are they buying it? The next step, of course, is natural, isn't it? How do we predict who's going to buy? Those are very important for doing forecasting and your supply chain and sales, especially. And forecasting is interesting because it is a little bit of a crystal ball gazing. And sometimes, especially if the people don't know what they're doing with the algorithms, they might choose the wrong algorithm and get a result and assume that's the right answer when actually it's not. Not all the algorithms work the same on the data sets that you have. So it's really important to have someone who's come from a statistical background to actually help you in when you're trying to get into the predictive analytics. And last is machine learning. Who's heard of machine learning here? Yeah. So I started using it in the 1990s. Everyone remember T9? I'm going to date myself on the old Nokia phones, the predictive text. Yeah. Oh yeah, what happened to that? LOL, right? That's what took over. So, you know, way back when in the 90s we had predictive text and that was a form of machine learning because it assumed what you wanted to talk about. Often, you know, when you're trying to write in slang or between friends, you maybe have, you know, maybe doing a disservice there, but it's gotten lots and lots better over time because of the natural language processing algorithms that are in use today. It still gets it wrong though, doesn't it? Not quite perfect? Yeah. So machine learning is really taking what you've learned in the history from business intelligence and the exploration from data mining and the predictive nature and then it puts action to that prediction, right? So you can do it on a large scale or you can do it in a small scale and I've got some really cool examples to share with you on that a little bit later. Any questions so far? Cool. If you have them, there's actually a question bank I think in the slider that you can go to if the question strikes you. Again, here's some more examples here to try and drive it home. It's important to actually look at these for a second, take them into your mind because I'm going to be asking you soon another workbook exercise. Who's experienced themselves some machine learning in those examples there? Anyone apply for credit lately? No? Okay, maybe that's a little bit too sensitive you have. So a lot of times in the insurance industry, bringing it to predict who's actually going to be a risk to you or not has been very, very important. And so they've had a lot of statistical people involved to understand what that risk is and put risk levels on individuals as they apply for an insurance package. Nowadays, and you've heard of UE as an example, everyone heard of that, these things are being done automated, right? There's a machine learning algorithm behind there that sort of looks at your details, ask you some additional information about yourself and then determines what your risk level is. That's taking the place of people doing that at the moment, right? So you'll see how this sort of thing is sort of coming into our everyday life. See a lot of people taking pictures. Do you provide the presentation afterwards? You do not. So take your pictures if you want them. Yeah? There you go. I won't be offended. Lots of pictures. You want me a photo bomb one? No. Look, yesterday was Halloween, right? I have to tell you. If you saw me yesterday, you'd probably not take me seriously today. But I'm an American if you haven't picked it out. So it's one of the holidays that I really get into and encourage my children. I had a zombie and I also had a Dred Pirate Robert running around yesterday. If anyone knows who the Dred Pirate Robert is, yes? Yeah, if you don't, go look up The Princess Bride and watch it this week and it's a great chick flick. Okay. What's that? I do have some pictures. Not in the slide day. See me. See me afterwards. I'll share them around my phone. You talk to my clients. They'll think I'm absolutely bananas. But that's right. You got to have good work stories, don't you? Okay, some example use cases. Right. So the first thing I'm going to do, I won't tell you a use case, but I will give you a reference and you'll want to probably take a picture of this next slide. So this was from Mackenzie. This report was put out in 2015. And what they did, they did a whole lot of research over a couple of years and actually looked at the value of this thing called the Internet of Things and Big Data. You know, why are we actually spending time talking about this? What's the big deal we don't understand? Right. So they went and did the research so that we could understand. And this is a fantastic reference guide for you. There's a small version and a big version. And it really goes into details industry by industry. I believe there's nine industries they look at and they try and examine where the benefits are going to come. What was really interesting in this particular piece of research was that the two areas that they really focused in was on streamlining operations, so operational efficiency and also customer interaction, knowing your customer a lot better. Those two categories rang true in all of the categories, especially in the smart city area where politics, excuse me, government and city government especially wants to know more about us so they can serve us a lot better. And that's where they see an explosion happening. There's a few data politics involved in that, but that's a topic of another day. Right. Some examples. Everybody can harken back to 2013. This year it didn't really pay off for Oracle, but in 2013 it did. I had a friend of mine, he was the head of the Oracle campaign from a technical perspective and he ran all of the systems and all of the analytics that provided them the information they needed to know to make the decision on the fateful day to change the camber of the wing. So what they had on their boat, they were pretty wired up. There's about 300 sensors on the boat itself. Each of the sailors had a PDA and a sensor, motion sensor. They could also hack in if they saw something a little bit different or weird. They do a little time hack on their watches. The watches actually, even though they were sponsored by TISOT, they weren't TISOTs, they were just knockoffs like, you know, this Timex I have with TISOT on the front because TISOT didn't have the watch they needed, which they actually built, which had the inputs they could put into the boat to actually do time hacks on when certain things happened. The sailors noted it. What else? They allowed them to basically feed 30, up to 30 variables into a real time system, which then gave them some ideas on what to do. Gentlemen, he's actually, he's a Kiwi. He went over to New Zealand, sorry, to America to run in the campaign and he's back now working for the Ministry of Health and using data and analytics to improve our health services at the moment. But this was a really interesting example and he tells me without the numbers coming back on terms of the sensors and the multivariate analysis on the current and the weather, they wouldn't have been able to say, all right, we're going to have some soft wind, change the camera of the wing, the boat's not going to tip over and go forth. And you know the rest of it, right? Biggest comeback. Not the second time. Will the health system work? Ask the current government. I'm not biased. I got no room to talk with my country back home. Okay. Zero. And that's another use case in social behavior with Cambridge Analytica. I could talk to you about that later. But today we're going to talk about zero. Everyone's heard of zero. Yeah, used it. So what we find in zero, and this, this is from a friend of mine Sandesh. He's the head of data and analytics in zero. And he gave this, this interesting case study a couple of months ago. It allowed me to recast it and let you know what they're doing at zero in this space. So at zero, what they're doing is you small business owners really many of them haven't come to the university or gone on a short course on, you know, finance and accounting 101. So they're accounting and the books that they have within zero are pretty shocking. Some of the statistics they have is, as an example, about many, many different categories for sales. One in five invoices are recoded from the sales account. Everyone familiar with accounting, you know, GL codes and stuff? Okay, cool. The challenge they have though is about every month, three million recodes happen in the, in just the simple sales category. And that means as a small business owner, I'm recoding, oh, it didn't mean to go there, it needs to go into this category or whatever. So it's a real, becomes a burden for the business owner who may not really care about running a clean set of books. They're more really reliant on the trade and what they're doing. So they tried to figure out, how can we actually help the small business owner in this regard? What can we do to help them? And so what they thought is well, just like predictive tax, account codes should be pretty generic. Let's see what we can do to predict what people may want to push this invoice into so they get it into the right GL code or the right category. And unfortunately what they found is because of this problem, the uniqueness of every individual business that's within zero, they couldn't just write one big algorithm that dealt with everything. What they found is they had to write little mini algorithms for everybody within zero. All the different individual businesses that are set up were very unique and different. And so they've trained the machine learning algorithm individually on a select group right now. They haven't run, excuse me, rolled it out worldwide to all the customer group. But that's an interesting fact that they found. And they also found that using an algorithm called, I won't tell you what it is, Secret Sauce, right? But it was better than just rules based. So in the old days we used to have rules. If this, then this, then that and boom, there's an answer, right? But with the algorithms it can be a little bit more dynamic, right? And so they found that to be double. So with the just rules it was about a 40% predictive rate. With the algorithms it was an 80%. So they decided to move forward with that program, right? They're helping us out, they're helping small business users. Ultimately what they want to do is not have any coding at all. Code free is where they're headed. And this is a little vision of what they see they're headed to after that. So, you know, as a business owner, I get this message and it does something for me, right? And as a business, it's on my watch. Yep, prove the payment off you go off to the next, next deal. So that's Sam Daesh. You can look him up on online and he'd be glad to talk to you about the smart stuff they're doing at zero. Another example is with airlines. So my business we do a lot with airlines. This is an example of what Boeing's done. So you think Boeing makes planes, right? Everyone thinks they make planes. They've actually turned into a data company. One of the things they've sold, especially with this Dreamliner, is a data program and a data contract with all the different airlines that they're servicing where they are now providing operational information and recommendations on the data that's being collected on these flying smartphones. There's about 140,000 data points or data collection points on one of these beasts. It collects it worldwide when it's in motion. All that data goes back to Boeing and they provide that service for fee, of course, to their airline clients. Airlines are also, so that's the manufacturer, airlines are also using information that they haven't been able to do previously. These things generate a heap of, heap of data every time they fly. But it's about a trillion records, okay? To store all that in the traditional fashion in a relational database takes a lot of time and money. And nowadays they have new techniques just to put it into what's called a data lake. In that data lake they can now run the algorithms on things, on the data and the flight information they previously could only aggregate up. So as an example I've got a customer who has taken that information, sliced it up or multiplied it by five times by taking smaller slices of a particular flight and understand the scheduling on time performance a little bit better in order to try and find where they can actually squeeze the time off the flights. When we're out there on the internet picking flights we're of course going to go for the shortest one. So five minutes here, ten minutes there means a hell of a lot of difference to this particular airline company's revenue. It's another example from flights. Again that was around operational efficiency and also with condition based maintenance. We're bowing. Anyone heard of the IDI? It's being looked after by stats? Yeah, one person? Cool. So there's a really cool story with this and it's around re-offenders. So what happened here and this was sponsored by Bill English. What they wanted to do is as we know department or even small departments have a hard time sharing data. Think about government departments bringing their data together. So there was an all government initiative that said right we're going to smash through these barriers and these tribal silos of information. We're going to create what's called an integrated data infrastructure. Stats New Zealand as the Switzerland of all the government departments has been told they're going to be the data steward. Right now they've got about 162 million facts in it and it's a beast. It's a 1.2 terabytes worth of data. And it's bringing data in from various different departments. One great use case on this and why that's important is because the more data sets you blend together the more context you get in the perspectives. Think about it as looking sort of at a three-dimensional object one way and then turning around the other corner and seeing the other perspective. If anyone's seen the OK Go videos where they sort of do those weird things it's one of those things. You need to see the full perspectives to gain the full idea of what's happening and that's what you can do with the IDI. So what got blended in this use case was government from MSD, sorry, data from MSD, data from corrections and also from justice. And by bringing those data sets together they could find and predict the people who are going to re-offend. And why they could predict it is because they found that after a parolee was employed we all know after the six-month mark what do we get as employees? Sick days. Sick days, exactly. So on six months and one day the parolees who then took a sick day were the ones that were more likely to start re-offending again. So that was an interesting thing that they were able to determine just by bringing those three different data sets together. So now the parole officers can intervene. Say, how you doing? You having a nice day? Do you like work? To make sure that re-offending doesn't happen because we don't want that in our community and it's very expensive on the government to deal with. OK, this is a local example that my team have done. So in Christmas it's generally a quiet period. So I give my team time to do hackathons and we've done one of these. This one is using publicly available data from Auckland Transport. Anyone tapped into that API feed? No? OK. So this is an interesting one. So they publish their data available to the web. You just have to apply for it. You say, look, may I please see your data? It's public data. They say yes and you can make about 35,000 calls a week, which equates to about every 10 seconds you can poll it. If you do that you get a general idea of where all the public transportation assets are at any given time. And if you then pull the other data feed they have around the schedule the buses and the schedule you can start to work out a picture of how they're doing performance wise. So I've grayed out the offenders but what my team have now done has done a bit of predictive and understanding which of the routes are always going to be late which of the stops are going to be late by the company and by the routes. And I haven't yet published this on our website. I'm sort of trying to work with Auckland Transport before I do that but maybe I do and maybe that's something that you want to use to find out your how your public transportation services going. A last one I've talked about all the great stories here's one that's not so great. This is an example for you just to just to gauge where you are when you put your ideas on analytic projects for your organization on maybe it's a step too far just yet. Natural language processing has been in works for many many many years right it's getting better and better and better. It's not quite there yet. This gentleman Max Douche tried to take all the Harry Potter books scan them so he did the optical character recognition threw it into his NLP engine and he asked it to write a book. Have you guys seen has anyone seen this before? Yeah okay so this is what the algorithm after ingesting all of those books in the context said this is a story that it wrote automatically. Okay that makes sense to you. Yeah not really there yet is it not really there. Yeah that's right I don't want to read into that first line but yeah so it's not quite there yet so when you're coming up with your ideas in a second or two think about that as well you know put your black hat on and say well actually maybe that's a little bit that's a bridge too far. Okay so the reason why I wanted to go through all the use cases was to try and give you an idea and sort of stimulate your thinking about in your own organization where you might find some opportunities right. So in your workbook going back to your workbook you will have the second page right list eight ideas where data can make smarter where data can use to make smarter decisions. So hopefully my use case is a stimulated your thinking on that and what I'd like you to do now in your own space is to write down eight ideas where you think in your organization data can be used to make smarter decisions. I'll give you two minutes for that. Okay if you could go back through on the margins of those ideas if you could categorize them based on what we talked about the four categories in analytics either business intelligence machine learning the data mining or the predictive analytics. Just write that down on the side and then we're going to learn a little bit about the audience and where their choices are. So back on the Slido site there'll be another poll open where I'd like to actually know where your heads are at where do you think the most value is by show of numbers in those areas. So if you've got for example a BI one mark a BI one you should be able to answer as many times as you have ideas. So if you've got it's an open poll so if you have eight ideas you should be able to go yeah I got two BI one data mining one predictive analytics and a machine learning to go please. Yeah you should just it's an open poll so you just just tick them tick them twice tick them three times we're really just looking at general ballpark. So what I think will happen is everyone will try and go for the sexy machine learning and forget about BI but we'll see how that goes. Hopefully not hopefully you've valued you know BI okay there we go yep some would say showing the answers influence the outcome but I'm not a physicist I don't work in quantum mechanics. Awesome thanks Oh did it it went away? Okay you don't want to influence the outcome John. John's a philosopher of ice schooling so we'll forgive him for that. Okay so the next topic I want to present is really try and get give you an indication of now that you've got these ideas what are the ingredients that's that's an interesting fact who are business intelligence still going strong pretty important okay I'm going to sort of give you some information about that would hopefully sharpen your radar in being able to discern which of the ideas that you wrote are actually going to be plausible or not okay and by doing that I'll just sort of share with you the the necessary ingredients uh-oh there we go back again I also will give a shout out to a colleague of mine he works for Dell EMC and he's also a lecturer at the University of San Francisco I met him a couple of years ago at a data conference in San Jose and he's been gracious to allow me to recast some of his ideas here either as a friendly knowledge share or perhaps maybe to sell more of his books I haven't figured out which so Bill's he's got a couple of books one's called Big Data the other is the I think it's the Big Data MBA both of those are powerful resource material for myself and my team and maybe you'll find them beneficial as well and he goes through this recipe and will only be able to in the time allowed be able to go through the first one really is actually understanding you know what do I need to actually get one of these analytic projects off the ground okay so the first thing you really need to do and it always starts with the business you know many times I get approached by lots and lots of people with ideas and say hey Devin come help me you know you do some startups come help me and sell this I've got a great idea and I really see yeah that's a great it's really it's unfortunately it's just heartening but inventors don't actually understand the business context and how to put those ideas in a way where you're gonna have commercial success so that those ideas grow and grow and grow so really you've got to start with that business initiative and often when those people approach me I go back to them and say okay how does it actually make sense in business how can you actually turn this into this idea into something that's commercial where it can feed itself has the cash flow and can grow and grow so I am really want to get it across the starting point for any of these isn't the cool thing that you can build it's really about why you're doing it okay and it's about the business initiative so here's the secret sauce these are the things that you need these ingredients here to actually make one of your analytic projects come to life right so you're looking at clear business ownership you're often gonna cross politics okay you're gonna get territories are gonna be drawn up information is power right you people will come out of the woodworks and try and shoot you down and you're gonna need that ally or allies to come in and help you bring people together have a nice informative and logical non-emotional conversation about why you're doing things really important to the business enough to give above the line most organizations are really working hard to do a great job and so the daily the daily task become huge and huge you know bigger and bigger so it has to be big enough to change the scope of the business and for business owners and users to actually provide the time to whatever initiative you're trying to get off the ground again a certain sense of urgency is important sometimes I look for what's called blunder funding so sense of urgency can be generated if there is for example a security breach in in a data set that may have personal identifiable information in there in that case that will absolutely strike the fear in the heart of any CEO because the minute you learn you lose the trust of your your customers your business is out of business right so sense of urgency is very important a big one is around data security and data sovereignty a compelling return on investment a learning culture you need a culture that's willing to adapt and sort of say yeah well this worked a little bit and then go the next next distance most of these projects get started with small initiatives they start as a proof of concept move into a pilot and then grow rarely do I see a successful analytic project starting as a big bang you might see that sometimes in the banks does it work I don't know they spend a lot of money I don't know if it works some of you who work in banks can tell me subject matter experts are really important so not the technical ones I'm talking about the business subject matter experts because they're the ones who bring that real life grass roots how is this actually going to work in our organization how are we going to change processes how are we going to change perhaps the people that we use and the things we're doing with the analytic outcome when we put it into motion a bounty of potential data sources and usually this isn't a problem these days and if you don't have one you can go get one anyone use flight radar 24 yeah a few of you fly around right so they had a problem in that there is there's enthusiasts out of Sweden came up with flight radar 24 they're limited in their resources and how many transponders they can grow in the community so what do they do they went out and sold they got little packaged prefab kits for the other enthusiasts around the world to buy for a hundred dollars and then of course their whole network grows worldwide so they didn't have a problem with that they got over their problem with having data sources not available or the capital actually fund that data quality is really important can't stress that enough it isn't until you get into one of these projects where you actually understand how bad your data is I'll tell you it's always worse than you think it is okay it's always worse than you think it is analytics friendly which means it's not pre it's not pre-calculated it doesn't have for example hierarchies it's really what we call low grain low density of information individual transactions or events and technical skills available to help you with that everyone who got a picture wanted a picture of the slide I got one three it's in the book it's in the book thank you very much it's in the book right good next exercise putting your analytic project ideas together alright putting? plotting right sorry about that okay so on your eight ideas what I'd like you to do now is pick three okay pick three write them in the workbook go to your workbooks pick three of those ideas and then to the person next to you I want you to tell them share with them why you pick those three ideas over the eight or more that you had on the other page if you only had one idea well you can say I only had one idea so I made it to this next page but really I want you to have a look at based on the ingredients you need for an analytic project filter through your list and pick three when you pick those three there's a few other categories for you to fill in in your workbook do the best that you can and then once you've done that look to your left or right pick a victim and tell them why you pick those three ideas please so I'll wait for the noise level to get to a dull roar before we move on to the next next exercise ten nine eight seven six five four three two one thank you everyone wrap up your chats great okay next exercise so just springboarding off your fervor for your ideas and your convincing nature to your partner to your left or right what I like you to do is then flip over to the next page of your workbook and plot those three ideas on that graph it's a simple xy plot looking at your relative ease and your against your return on investment again this is all just your judgment plot those three ideas on the graph just a little scatter plot just a plot, a dot maybe write the name to remind you or one, two, three ABC Z, Y, X if you like to go the other direction okay if you've plotted those pick one that you want to do when you pick that one I'll ask you to go back onto your smartphones and type in the working title for that idea what we're going to do now is find a word is going to form a word cloud from those those working titles that you type in and we'll see what's popular so pick one go to the Slido app and write in the working title just type it in as best you can brevity is best so if you look up on the screen you'll see where the commonality may be and the size of the word tells you the usage really just if it's used more than it's bigger fantastic, look at all these ideas just since the only thing I didn't do was bring in the venture capitalist just at the door all the VCs that come in and harvest these good ideas and work with these bright individuals in here yeah great awesome look as we wrap up I'd like to start by first thanking you for your appreciate your and appreciate your contributions and your time and the amount of effort that you all have put in today to hopefully I love thinking of ideas, I love brainstorm so hopefully it's been a really good experience for you again I do appreciate your time and your patience and the initial technical difficulties that we had my intention today other than just to entertain and have a little bit of levity in the day was to really help you understand why we need to be data driven today what's the importance of that also to help tune your analytics radar so in your own organizations or in your community you can have an eye out for ideas where hey look if we do this a little bit smarter use this data maybe we can do some analysis and get a better outcome right and so hopefully tune your radar on those projects that are actually going to be feasible and move forward lastly what I wanted to do with you today is actually help you be a bit real so plot those ideas against the relative return on investment and also the ease of getting it going so that actually you can be your own personal sanity check on those ideas and then maybe to take that idea bring in another person and move that idea forward so again I appreciate your time today I hope you learned a little bit I'm going to be here outside afterwards to answer any questions actually I think we're going to take a couple of questions first yes yep cool so if you have any questions how much time do we have for those two or three minutes okay great any questions from the floor at all I'll ask one first who's doing the tough motor this weekend right when we're on the obstacles together help each other out that's what it's about cool um... yes sir there's a microphone okay cool thank you right cool thank you for your question I can't speak for the department of corrections on this one but I do know a few people there on that and I think it's one of those what they tell me it's actually it's because of the the cohorts that they're involved with so being able to say you know are you okay on if they do take that sickly is important but also understanding their cohorts could really it's they fall back into the compute community that they used to operate in from what I've been told so I don't I don't work for the department corrections I I can only tell you what I've been told it's really about that's one indicator that they may reoffend but also there it is a campaign it does take a community and there's different social services to help them with that to understand to prevent them from reoffending so I don't know if I actually I danced a little bit I don't know if I answered your question sorry about that anyone else depends on the country depends okay so new zealand has some interesting laws I won't purport to be a lawyer or know them intimately right but there are some privacy laws in new zealand which are probably a bit more strict than other places such as the u.s. so I can give you an example of the u.s. yeah and so that may or may not relate to what's happening in new zealand on the privacy laws everyone knows of the campaign result a few months ago and they've heard about how analytics were used to influence decision-makers on that yep cool so in america you leave a digital trail in fact john can you go up and show that tab with the um... those analytics so when I started this comment that's a great conversations starter when I started this event today I asked you if I could learn a little bit about you I asked for your permission to do that and most of the organizations you still have to ask that permission but once you grant that permission then um... it's not one of those it's um... might have to go over there and fiddle with it there it's not on the slido it's not on slido it's the demo EIT one it should be as a favorite anyway what he'll show you is um... when I asked you for that and you gave me permission I asked you to go to the demo dot EIT in zed website by doing so you left you started leaving a digital trail so that website captures your browser information and actually captures the operating system that you have on your phone okay so if I were a marketer I'd use it doesn't capture who you are doesn't capture any personal identifiable information about you other than this is a browser and this is the phone that it's using right and the um... the release of the operating system so let's say I worked for apple and I also looked in the poll that we had earlier about where did you sit you know did you sit because you wanted to speedy exit did you sit because of friends or or did you sit because you wanted to meet new people or was it because of the view well I can actually use that information and tie that to what I know about your personal iphone and then target a message to you so let's say that you're socially influenced by others you like to be part of the group perhaps you've got an older phone I might send a message to you saying hey did you know that so-and-so or your friends or these people like you actually have gone to the latest version of the apple iphone or samson I can target that message specifically to you but because I ask for permission I can do that I can harvest that information most of the websites that we're on today especially the social media ones and I gotta tell you I have a couple of wonderful people from china who used to work in these internet giants harvesting this information and they are so scared about what they they're so conservative about what websites they join because they used to do that as a as a trade is get that information in the stories they can tell you what they found out about how people using your phones or where the using everything else talk about no data privacy absolutely when you sign into is that we chat whatever they capture everything about you and those without an ios are even worse all those applications work together and bring all that information together to the application vendors right so be sensitive be sensible do your research it's a great question if you grant permission they'll give everything from you right it's a trade off between all i like your application you do me a service they're collecting information about you all the time especially location based information which is really really important back on the topic of trump and brexit today say that maybe i did uh... so what happened there is a collected information and they bought it is cambridge analytic uh... from facebook and other social media sites they absolutely entitled to that they bought it it's there they can use it for their advantage and they targeted their messages to people who would more likely be moved who actually hadn't made a decision now where it's interesting is that for u.s. campaign uh... that information was processed in in um... in in england right in a data center in england is a professor in new york city who's actually english and the rights in in england is if if you're using my information i mean and i want to know what you're doing with that i can ask you and you have to tell me what cambridge analytic has said is what actually you're not in england so you're not it doesn't matter but the point is the data was processed in england so now there's a lawsuit pending if any of you are in the legal profession i can tell you will make a mint on prosecuting and and this whole idea about data sovereignty and individual rights for data it is an emerging field absolutely absolutely good question did you find the website john computer died what you don't know i left home without my laptop cable in my battery life is out this poor planning prevents this poor prince good performance anyway right one more question that's it one more quick question everyone wants to be here even wonderful thank you very much for your protest thank you very much