 I won't be fiddling with my phone as I speak, it's control in my lap. Just in the case of being not speaking and texting. It's a great pleasure to be here today and I've been really looking forward to this and I look forward to chatting with you afterwards. I think I'll purposefully leave open some big questions at the end but then I'm sure you'll want to ask. I've been around big data and privacy and ethics and all of that. At the moment I head up a centre called Insight which is a large research centre in the area of data analytics. It's a very young centre, we're just about a year old at this stage. It's funded by Sirens Foundation Ireland and it brings together data analytics expertise in UCD, Dublin City University, NUI Galway and UCC and we also have some smaller interactions with a number of other institutions as well including the Royal Irish Academy. The significance of it is that there's about 250 data scientists working in Insight today which is a really significant critical mass for this type of expertise and I'd say that outside some of the largest organisations, Google and Facebook, you won't find that critical mass of data science available so it's to the credit of Ireland and the government and Sirens Foundation Ireland that they've recognised the potential of this area and invested so wisely in it but then I wouldn't say that. So let me start by talking a little bit about data. You know almost everything we do causes data to be recorded and stored somewhere. We might be listening to music or watching a movie, we might be exercising, we might be shopping, we might be sleeping or even taking a shower and probably somewhere a data record has been created and we might think about what could be possible from all of that data. What can we do? What should we do? What are we doing with all of that data? And people talk about big data and you wonder well how much data is there out there anyway and to really understand this you need to think in things called exabytes and exabytes are very large amounts of data. 10 to the 18 bytes, what does that mean in real terms? Well if you took the US Library of Congress it's about 20,000 times all of the printed material in the US Library of Congress that's an exabyte or all of the words spoken by humans, ever, ever, that's about an exabyte. Well we're creating that amount of information about every six hours. So it's a stunning amount of information that we've created. And one of the things that has happened over the last few very short years in fact is there's been a paradigm shift that used to always be about the algorithms. So I'm a computer scientist when I went through my training it was all about building smarter, faster, more efficient, optimized algorithms. And feeding those algorithms with data, the data was kind of secondary. Now we're beginning to realize that actually the data is the really important piece here and sometimes even some simpler algorithms that we might have looked at twice in the past begin to work extremely well with large amounts of data and the right type of data. So there's a shift and data is becoming really, really important. The algorithms are still obviously critical. So how did we get here? I'd like to talk, to go back a little bit. I suppose this is a bit farther back in your story anyway, but you know when we think about the first large supercomputing, the very first order was the CDC 6600. It weighed over 5,000 kilos. It came with about 400K, 480K of RAM. It cost in today's money about 60 million dollars. And it was able to do 3 million flops. What's a flop? It's a floating point operation. Think of it as a basic mathematical operation. So you've got about 3 million of those a second for your 60 million dollars. That was the start of really the computing revolution. And Gordon Moore, who was one of the founders of Intel, he noticed early on that there was something strange happening in the computer revolution because about every 18 months we were doubling the power, essentially, that we were getting from these devices for the same price point. And that became known as Moore's law. And Moore predicted that this might last for about 10 years. Well, it's lasted for nearly 50 years on this stage. So we're seeing a doubling of computational performance every 18 months or so over the last number of decades. So much so that if we go from the 1960s, we were getting about 0.05 of these basic computations per dollar. Today in my phone, I'm getting about 8 million of these computations per dollar. My phone is thousands upon thousands of times more powerful than that supercomputer, which weighs thousands and thousands of times more than my phone. If the automotive industry had moved at the same speed, in the same way, we'd be driving around in cars that would cruise comfortably at a million miles an hour, and we'd get about a million miles per gallon of diesel. And it'd be cheaper just to chuck our car away at the end of the day rather than park it overnight in the city garage. So clearly the motor industry didn't develop at the same pace. But that's why we've gotten to where we are, that these computational devices that we're so used to are now so incredibly powerful and so cheap for that power that a whole new type of world is possible when we start to knit these together and connect these together. And often people talk about the sensor web because what we're finding is that things that were kept inside of these machines, inside the virtual world of the internet, it's now leaking out into the physical world because our physical world is being equipped with sensors. So we're now able to read a lot of what is going on in the world around us and influence what is going on in the world around us. And I'd like to take you on a bit of a whistle-stop tour of some of the things that I think are really interesting about what's happening in the sensor web, particularly in the area of healthcare and how that will be disrupted over the next number of years. And we look at this from the perspective of the mobile phone, which we think of, well, most people still think of it as a phone. It's very far from being a phone. It's not even a very good phone anymore in my opinion. It's a great internet device, the great data device, but it's an even better sensing platform. And when you think about all the different sensors that exist on these devices, which we carry around with us every day, light sensors, microphones, GPS location, digital compasses, and so on and so forth. And then we realize that more or less they're always in power. So power in sensors used to be a big challenge if you're deploying water quality sensors in a river system. How do you keep those sensors powered all of the time? Well, we keep these sensors powered all of the time. And they're always connected to high-speed data pretty much, and they're mobile, and they know where they are. So that's a very powerful idea. And it promises to completely disrupt many different aspects of the world in which we live. And I think healthcare is a good case study. This is Eric Topol. He's a cardiologist in the US. And a couple of years ago, he was on a plane travelling across the country, and another passenger was taken ill. And Eric had a small ultrasound device connected to his iPhone, who was able to diagnose the patient was having a particular type of heart attack, was able to do the right thing there and then on the plane, and the patient recovered when they landed subsequently. But that just points to the disruption that's happening. An ultrasound scanner connected to an iPhone. Who would have thought that would be possible only a few years ago? But let's have a look at some of the things that are possible today and that we probably are using in apps that we have. For any of you who are into exercise, fitness, running, swimming, walking, cycling, many of us who are doing these activities are tracking our activity on an app. I use an app called RunKeeper. Every time I go out, I wear one of these jaw bone devices, which is tracking my activity. It knows not just how much I exercise, when I exercise. When I'm wearing a heart monitor, it knows my cardiac rhythms during my exercise. Where I go, the altitude, I go up or down, who I'm with. Times of day, it knows my gender, my weight, my age. Imagine that information going back to my GP. Not just a periodic once a year check-up, but a long-term track of my activity information going back to my carers. I think that's a very interesting idea. Imagine how valuable this is to the likes of Nike and Adidas in terms of developing next-generation health equipment. People are using apps to track your sleep. You can get a free app for your phone. You put the phone under your pillow, and it tracks your sleep based on movement during the night. Devices like this also track my sleep. I thought I was a poor sleeper. I'm sorry, I was kind of a bit obsessed about my sleep pattern. So I started tracking my sleep. I'm not really that poor a sleeper at all, which made me feel a lot better about my sleep patterns, which got me to sleep a lot faster. So now I'm a really good sleeper also. But we're tracking this over long periods of time for free using our phones. Who would have thought that a phone was tracking our sleep? You've got some companies now that are even able to track our mood and our focus, unfortunately, to wear a headband, that will change in time as well. People are tracking their food and their nutrition. How is that going to change health care when doctors understand on a meal-by-meal basis what you're eating? I have a heart rate sensor on my phone. I put my finger over the camera. It tracks my heart rate in real time by looking for tiny color changes in my skin as it shines to flash on my finger. As my heart beat beats the color of my skin as blood flow changes and it tracks that. I can point the camera at my face. It'll measure my heart rate based on tiny color changes in my face. Again, there's no extra sensor on the phone. It's just the camera, but we're able to track people's heart rate and where people are able to do that not just once or twice a year when they go for a check-up, but maybe before exercise, during exercise, after exercise. Here's an interesting one. It's not a demarcage yet, but it is an experiment to see what you could do with the speaker on the phone. This is a mobile spirometry. It's a lung function test. The patient here is just blowing into the speaker on the phone and it's able to do a standard clinical lung function test within about 5% of clinical accuracy. That's good enough for home usage for people who are suffering from asthma, for example, or other lung issues, just by blowing into the speaker on your phone. All for free, all as part of the device that you carry with you everywhere. One of the things I'd like you to think about is how might this change healthcare? I think healthcare will be slow to change, but I think it'll be us, the patients and the consumers that drag it with us. So might we be going into a doctor with our apps, with our data, loading it up to the doctor's system? There was a BBC did a horizon documentary on this recently where they talked about, instead of going away with the prescription, maybe you'll go away with a new app on your phone because your GP wants to look at some particular aspect of your activity or measure some particular constraint and there's an app available to do that. So go away, let this app monitor you for a couple of weeks, come back and we'll take a look at where we are. I think that's a very real possibility. So we have access to this data now, it's valuable data. We'll come back to that point in a while, but it's data that could fundamentally change the healthcare system and the quality of healthcare that we get, so that instead of periodic once or twice a year check-ups, we're bringing long-term, longitudinal physiological data with us. One of the really interesting things though is that it's not just about individuals carrying phones and sensing their individual activities. We can scale this up across entire populations. These phones are going to be equipped with more and more sensors, so we scale it vertically, we're measuring more and more about individuals and of course then they will scale horizontally. We can start to connect up these sensors into large networks that would be far too expensive to build by traditional needs. This has led to this idea called participatory sensing where individuals form a sensor network which can start to read interesting things about the world that we're in. I'll give you my favorite example of this. It's a U.S. company called Propowder Health and they produce a sensor that goes into a standard rescue inhaler, it's about a dollar sensor and all it does is it knows when the inhaler has been used, but it's connected to the patient's phone. So it knows when the inhaler has been used, it knows where the user was when they used that inhaler because the phone knows where it is. So why is that useful? Well, imagine connecting that up so now you're the city with thousands of people using these inhalers and the interesting thing about inhaler use it tends to be heavily correlated with air quality for example. So now you have an air quality sensor network by accident almost that would cost millions to build by traditional means where it means putting physical sensors on land posts for example all because of a dollar sensor that was designed originally to tell you when you were running out of medication on your inhaler and when you needed to restock. There are all sorts of examples of how ideas like this connecting sensors together are transforming many traditional industries and one of them is the mapping industry. It used to be that Google bought satellite companies so they would be able to build accurate maps of the world but now we have companies like Waze which not surprisingly were subsequently bought by Google but Waze built their maps by giving everyone a free smartphone map that just tracked where you went and as you went about your business went about your commute essentially an empty map started to be filled in by the trails that people would take through the world and those trails were reinforced so that you could start to confidently assert that they were roads or pathways and then you would invite people and everyone doesn't have to do this but enough people were willing to mark down the names of the roads and express points of interest for example and all of a sudden you have real time maps of the world developing out of the activities of many millions of individual users without expensive mapping software meaning to be developed and not only that because the phones are being carried with you it doesn't just give you a static map of the world it actually gives you information about traffic congestion so Waze knows that there's a lot of traffic in this part of Manhattan at the moment for example because the phones are going slower that they usually are at this time of the day so there's a lot of benefits coming from this sort of sensor webpinking and it's all about turning people into sensors essentially and inviting people to participate in these sensor networks and it starts with lots of messy data but we now have the algorithms to clean that data up and to discover real genuine insights such as poor air quality in this part of the city don't travel that way to work this morning take this route or traffic congestion is particularly bad at the moment down there go this way it's all based on the smartphone as a sensor platform and the willingness of people to contribute their data it's not just what we're doing in the physical world we're also used to data being generated in the online world and this rather unfortunate phrase minding the data exhaust as we pass through the world we leave an exhaust trail behind us which can be mined and there's very valuable information in that it's the pages that we look at the time we spend reading those pages even the mouse movements that we make as we use the browser for example our search queries and companies like Google of course have been particularly successful at minding page relevance and user reputation by the links that people follow and they have used that to build better search engines and to deliver better advertising but what you might be less familiar with is the importance of search logs for other types of applications so a group of Microsoft researchers recently speculated that by looking at the queries people submit to search engines we may be able to discover unknown interactions between drugs I think how is that possible well they looked at queries going through the Bing search engine and they were looking in particular for people who queried for peroxateen and then there was other people who searches were for pravastatin and then they were also looking for people who were tending to query for both peroxateen and pravastatin now the hypothesis here was that the combination of these drugs may result in hyperglycemia so they looked to see whether there was a higher prevalence of hyperglycemia symptoms being searched for by the people who were taking only pravastatin or only peroxateen versus the group who seemed to be taking both pravastatin and peroxateen and to go long story short they found a significantly higher hit rate of hyperglycemia terms from these searches which confirmed or at least pointed to the possibility of an interaction between both drugs leading to hyperglycemia and as I understand it this has been clinically proven so there's a sense that mining search log information can actually now start to help public health officials get a much better handle on previously unknown drug interactions which is a very expensive thing to try and do normally during normal drug trials similarly you have seen over the last few years that Google queries have been used to track the outbreak of flu now recently this has been brought into question because they didn't get it quite right over the last year or two but there's a strong correlation between people searching for cold remedies and flu remedies and the outbreak of flu location by location and Google flu trends is an app that will actually predict flu epidemics breaking out in different parts of the country based on people's search queries we're also now starting to dig into the reviews that people leave online so you'll be familiar that Amazon kind of tracks what we like and dislike and what we buy and makes recommendations to us but we can now start to actually read our, systems can read our reviews and identify our opinions from the natural language text that we're providing and recognise that we like a MacBook Air because it's an intuitive device and it's got a great display and image quality is really excellent but it's a bit pricey and actually extract that really high quality information we did an analysis recently mining hotel reviews across a number of US cities and we found very different opinions being expressed in very different cities in Chicago the things that people were most happy about were location, staff, the quality of the bed and the service and the breakfast versus the things they were least happy about the noise, the carpets and the elevators so this is providing hotels with real information about areas that they can improve on and it's picking up things that matter to travellers that are rarely expressed in the description of the hotel who to think that people are interested in the quality of the carpets but certainly hotels probably aren't paying as much attention at least in Chicago as they should the carpet quality here's one that Ayman might be a little bit familiar with we mined sentiment for during the last election from what people were saying about politicians on Twitter so we're able to say who's been talked about the most that's the size of the head here what are people saying about them is it positive or is it negative and this is changing in real time you can see switch, shifts and sentiment in real time and shifts and volume in real time, thank you so yeah, go on to the next one, thank you so this is really about data driven everything that I'm starting to talk about here we've seen some healthcare we've seen mapping we've seen online sentiment and opinion mind so much further than that and entire disciplines are being transformed by big data thinking everything from social sciences and humanities to political science and journalism this is Google's engram viewer it allows you to look at the relative popularity of different pairs of words in books dating back to the 1900s so Google has basically scanned most of the books that are available in print you can actually look for shifts in word usage over long periods of time this was a particular paper on the expression of emotions in 20th century literature and you can see shifts in the ratio of joy versus sadness words over particular periods that correlate with major events at those times instead of joy, sadness, fear and disgust the rise of terrorism and the increased usage of fear words, for example in modern literature who would have thought that this type of research would be drawing on Google logs for example in Google data and I just took a sample of some recent papers that have been published to highlight quite how many disciplines are being transformed by data analytical thinking we have the evolutionary dynamics of language and environmental issues and ethics cultural studies so on and so forth so I think what I'd like to finish on and sort of lead into discussion on is that I think for sure the big data revolution is here to stay and it's here today so the genie is out of the box data is now being generated on almost everything that we're doing and companies and organizations are starting to wake up to the power of that data it's very transformative it's very disruptive and people recognize that there are huge opportunities to win from that data not just in business but for societies it's all about helping people to make better decisions helping individuals and families and communities to make better decisions about their healthcare where to send their kids for school help governments to make better decisions about investment for example it's not just about helping large corporates make better decisions about who to target and how to extract the most value from their customers and that being said I think there are a number of big issues that we still need to resolve as a society and what I haven't spoken about is the privacy implications of big data data is being collected on all of us all of the time and the one thing is certain is that we know very little about what is being collected and we certainly have very little control over what is being collected even over every week there's another story on the dangers of misuse of personal data in particular last weekend it was Facebook with their emotion study and the ethics of what they did there so I don't know if people have seen that but it's pretty clear that we need as a society to figure out some of these things and personally I think that part of the solution is going to be giving individuals ownership of their data so that we can control how our personal data is used there's a phrase that's often used when you talk about companies like Google and Facebook where you're not paying for their services well if you're not paying for their services you're not the customer you're the product and that's a very accurate description of what's happening we're selling us to advertisers we're not Google's customers we're Google's products and we need to understand whether we're satisfied with that and there may be because we may decide that the quality of service I get from Google is good enough and I'm happy with what they're doing with my data but equally I might decide I'd rather pay Facebook 20 or 30 dollars a year and keep my data in my control and give everything to Facebook these are the sort of questions that individuals need to start asking that different people will land on those answers in different ways but ultimately I think a big part of this is giving individuals control of their personal data and developing the privacy and ethics framework around all of this that will allow us to live in a big data world more responsibly that we're benefiting from a healthier, safer and fairer world as a result of this technology so I'll pause there and I'm happy to answer any questions you might have