20 January 2016

Big Problems? Big Data. – Matt Harding

Matt Harding shares how Big Data can help solve public policy challenges like rising global energy demands and our expanding American waistline.

[MUSIC] This moment in time is defined by our need

to solve some really big problems-from environmental challenges, to terrorism, to healthcare. We are dealing with really complex problems. In fact, these problems are so complex that sociologists call them wicked problems. They are problems that don't have a single solution. They require a complex interdisciplinary line of approach in order for us to even get a handle on how to actually solve them. But at the very core of all these problems, we are fundamentally dealing with behavior change. We need to find ways in which we can change human behavior. Now big data is a tool that we have that allows us to make better decisions. I am not afraid of big data, and I invite you to embrace it rather than be afraid of it. But first, let's think about the digital universe in which we live today. It is an expanding universe, and it's expanding at an astonishing rate on a number of different dimensions. So first, think about the data volume. We used to measure our data in megabytes. Now we're measuring it in brontobytes. [LAUGHTER] Over the next two days we are going to generate more data than our entire civilization has generated from its very beginning, up to about the mid-2000s. This is astonishing. But it's not only volume. It's also about the velocity with which we create data. We used to wait for months or years to get new economic data that allowed us to do forecasting. We no longer do forecasting. We do "now-casting." And thirdly, the variety of the data has changed tremendously. Our data used to come in this nice little spreadsheets in Excel. Now we are all creating data, so I hope that tonight, as you are tweeting and taking pictures

of me, you are helping increase this expanding digital universe. So data is here and it's here to stay, but where does the value come from? What we do today with data goes well beyond its size. We can create a lot of connections between data sets that may seem completely unrelated, and by doing so, we adapt to this problem, and that's where the value comes from. One of the things that we really focus on at my center is what I like to call the pursuit of triple win strategies. Data is the kind of connections that big data allows us to look for. There tremendous opportunities to actually find solutions to some of these big problems that benefit consumers, are profitable for firms, and also promote some sort of social good that we all care about and can, can stand behind it. So let's talk about a couple of examples of the kind of work that we do, and let's think about something as simple as-as going to the supermarket and buying food. The average supermarket has about 50,000 products. Overall in the U.S., we are buying around 1.1 million products. All of these products are barcoded. They're tracked. And that's a good thing because it enables us to know where these products come from. We know that the food is safe. We know that the food is local. It's organic. It's-shows us about the things that we want to know. But at the same time, when it comes to food we are facing some really major challenges. Over 35% of Americans are obese, and over 15% are hungry. How is this happening in the world where we have such an abundance of choice? That's why it becomes so important to go beyond the data and think about how we can combine the data with the work that we do in behavioral sciences to understand how people make choices, and then partner with outside organizations, with retailors, with companies out there in the real world, and actually try and find ways in which we can encourage people to make healthier food choices.

So we particularly focus on low-income households and households on food assistance programs and try and encourage them to make better choices when it comes to food. Another area where we're seeing a tremendous amount of innovation is in the electric grid. The electric grid is undergoing its most significant transformation since Thomas Edison. We are moving from an analogue system to a digital system. It's what you heard of as the-the smart grid. So let's think about what this actually means for your house and many of the homes out there, because now we have installed millions and millions of smart meters. So we go from a system where the meter that you have on your house may look something like this-and really all it did is record the amount of electricity that you are consuming-to a much more complicated and sophisticated technology that's embedded in a smart meter. And what a smart meter does, it enables two-way communication between the outside world and every single device and appliance that you have in your home. Imagine the world of tomorrow where your HVAC system is no longer going to break down in the middle of the summer when it's really, really hot. In fact, what's going to happen is that all the sensors that are embedded in the HVAC system itself are going to communicate through the smart meter, with smart programs. They are somewhere else, and two days before your HVAC system is about to break down, the contractors are going to come at your door, and it's going to knock on your door and it's going to say, "Let me fix it before it breaks down." We are well on the way to actually achieve this. So let's go back and think about behavior change. Here's a fact from a recent study by McKinsey: 16 to 20% of the energy consumption in our homes, in our residences, could be saved if we were to engage in some very, very simple behavioral change. What I'm talking about here are things like turning the lights off when you leave the house, washing your clothes in, in cold water instead of-instead of hot water, and we all know that those things save, save power. It is not a surprise to anyone. So why aren't we doing all these things?

Clearly, it's good for-it's-it's good for our pocket. It's good for the environment. We are not doing this. Well, it turns out there is something about human nature that even when it comes to these things that are-are completely obvious that we should be doing them, that we actually need a nudge in order to do them. So, one of the projects that I did with a utility in Chicago, it was basically about creating a website where consumers could go, and they could choose a goal for their energy efficiency, for energy savings. And what big data enabled us to do is to provide each household with personalized suggestions on how they can actually save energy. So over an 18-months period, households who choose realistic goals saved over 10%. So that is astonishing because it didn't involve anything more than a simple website. But as we're thinking about behavior change, and also the impact of data, we constantly have to pay attention. And data, what it does also today, it creates an information avalanche. So it's hard for us to pay attention. I know it's very hard for you not to check your Facebook account right now, but we have to do it. At the same time, the average American household spends about less than nine minutes thinking about energy. So how are we going to get all these people to pay attention to energy and to reduce their energy consumption? The answer doesn't lie in more websites. I think the answer lies in engaging with smart technology. We are now at a point where deep data and big data allows us to create devices that are very smart. They allow us to-to automate a lot of the behaviors that we would normally have to do ourselves. So rather than having to-to pay attention to everything that is happening, we can have

smart technology that is going to pay attention for us. So let me show you some-a quick result from an experiment. A group of households faced higher electricity prices in the afternoon. It was hot. Prices went up by a factor of 10. Those households were then randomly allocated between two groups. One group which had access to a website alone that, essentially, required them to pay attention to prices and their own demand all the time. And another group that was given a program of a thermostat where they could enter their preferences and the thermostat would respond alone. Households who thought they're going to be able to pay attention saved about 10%. Those who had the thermostat that actually paid attention for them saved over 60%. So if there's one thing I want you guys to do today, or tomorrow, is to actually go and buy a smart thermostat. So as we move forward, we are moving into a world of connected devices. Every single device around us collects data, it communicates with all the other devices. We can try and pay attention to all of this. We are not going to be successful. A much better investment for us is to actually invest in technology that automates a lot of these behaviors. And we are still very much part of the process because our preferences, our values, the things that matter to us are things that can be entered into these devices, and these devices will automatically respond to our needs over time. Now, big data is also something that connects everything that we do at Duke-from medicine, to social science, to engineering, we all engage in big data nowadays. And it is also something that allows us to connect students, and faculty, and researchers, and we really add that depth to the volume of data that is available out there. Because data allows us to find out the facts. It allows us to develop new policies and new programs that help solve some of these big challenges out there, and it also provides us with a means to formally evaluate and measure

their effectiveness. Thank you. [APPLAUSE] [MUSIC] [END RECORDING]