 Live from the Hilton at Bonnet Creek, Orlando, Florida. Extracting the signal from the noise. It's theCUBE, covering Vision 2015. Brought to you by IBM. Welcome back to IBM Vision, everybody. This is Dave Vellante, and this is theCUBE. theCUBE is here at the Vision conference. We're talking about governance, risk, compliance, sales, performance management, but we're going to kick this off with a really interesting session. David Haas is here, he's a cyclist, and Jean-François Pouget is a chief architect with IBM. And gentlemen, first of all, welcome to theCUBE. Thank you. So this is a very interesting story. David, you are actually going to be performing. We have, as I've been mentioning, IBMVisionGo.com is our website, and we have a third channel, which is the Ultra-Cyclist channel. And David, you're going to be riding this afternoon, but you guys have connected up with a little Internet of Things example. So David, why don't you set it up and then, Jean-François, we'll talk about what IBM provided here. So I'm performing a racing in the race across America. It's a 3,000 mile bike race non-stop. Pretty much ride the whole time, sleep for very limited amounts of time. And the finishing, the winning time is usually under nine days. So you have to be on top of your performance to continue to ride your bike that long. So we're riding anywhere from 22, 23 hours a day, trying to make very good decisions about what's going on. And so with the analytics that IBM is helping me with, I can measure all kinds of different things using the Internet of Things. So we'll measure heart rate, power, speed, cadence, some standard things we use in the bike, but then we're going to be able to take some of that and expand beyond it with using some weather data with the weather company involved and use wind conditions to make better decisions in the race. So it's a really cool idea and I'm hoping that it improves my performance in the race. So when I started blogging years and years ago, I perfected the art of typing while I sleep. Can you ride your bike while you're sleeping? I've fallen asleep on the bike, I haven't. That can be quite dangerous, luckily for the helmet. So Jean François, essentially you're instrumenting David. Is that right? We start with instrumentation. But to many people, analytics means looking at data, finding trends. We want to move farther, we want to move to decision making. So we will collect data, we will use that data to predict his performance in the coming hours and days. We will get weather data and especially wind, trends and direction because when you cycle, one of the key factor impacting speed is whether you have front wind or tail wind. And with this, we will help Dave decide when to rest because the key decision, as he said, he will run 22 hours a day or so, have two hour rest only per day. And the key decision is when does he should take this two hour of rest. And we're helping, proposing alternative and finding the optimal time for resting. So a lot of that's predictive. Predictive and then prescriptive. So the predictive is predicting weather conditions. So that's a weather company. And then predicting how fast he will ride across the road. And then prescriptive, which is helping him choose when to stop. So going from the insight from the prediction to the action. So the primary parameter for you, David, as a cyclist is rest, are there others in terms of how hard to go? Are there equipment issues like when to stop and change? I'm thinking about a Formula One racing, you know, tire wear, things like that. Or is that really primarily come down to when to rest? The biggest issue with the race is obviously if you're on the bike moving forward at a fast pace, that's the key. So if you can do that nonstop, which you, you know, my fitness says, you know, so in the past I've used fitness and luck. And with IBM, we're using some foresight and beyond that, we're taking the ability to make decisions. So my power is going to decrease the whole way across the country. I mean, obviously, but our goal is to have it decrease the least amount. So, you know, we're riding across the country at a power level that I can maintain. Then I'm going to take some rest based on the prescribed conditions. My power should come back up and then it's going to continue to decrease. But it's trying to keep that power at the highest level we can while we're riding the bike and get to the finish line the fastest. So using that information, as opposed to just using it intuitively and saying, wow, I feel terrible today. We're going to be able to use real information and real data and real performance that we're using out on the road to make very good decisions while racing. And during this race, I have a crew that follows along and helps me with those decisions. So they give me water bottles. They basically are behind me the whole time while I'm racing and they'll have a dashboard in the vehicle that's going to monitor all the information that IBM has put together and what JFP has put together for us and will help them make decisions. So it'll say you should continue to race because of all of the conditions or the conditions aren't looking favorable and your power isn't favorable at this time. So it might be a good time to consider resting at this point. And how's the comms? What's the communications? Is it verbal? Do you have an earbud? Is it wireless? So with my team, I will have an earbud that I wear that is a Bluetooth wireless earbud. But all of the data is captured through a Garmin device on my bike, which then will go through an Android phone and through the cloud to the IBM, basically the iPad that's in the car. And is the primary communications back to the rider in this case verbal or do you have your own little Garmin dashboard as well? It would be verbal from my crew giving me information. Okay, so it's your info out and then verbal back. Yeah, and I'm watching my own Garmin to maintain proper power and things like that. So I am watching a device that's telling me what my power is and how fast I'm going and things. I'm not getting, you know, I'm going to feel the wind direction and boy, it feels windy, but I don't know what that wind is. Or even better, I don't know what that wind is, 25 miles up the road where with this information, I'm going to know that we're hitting a headwind, which in your kind of that question might lead to saying, we're going to have a real good headwind and I want to be on my most aerodynamic bike that I'm using because I'll have multiple bikes. So to fight through some of that or take advantage of the wind, I may change what piece of equipment I'm riding on to do that. So we think about weather being very unpredictable, but there's a more predictable aspect of the weather that you're forecasting, which is wind. Actually, you're not forecasting, you're partnering with somebody to forecast, right? But that's a big assumption in the data. Do you worry about the sort of garbage in garbage out and how do you manage that? The wind prediction, we don't expect it to be accurate for more than two days. So we will use it to basically fine tune when is the next rest, in the next 24, 28 hours, when is the next rest? But for the long-term strategy, we can't use wind prediction for the full road. It won't be, as you said, garbage in, we would mean garbage out. But I want to come back. There is a study done on a Ram runner a couple of years ago that showed that perception, the way the cyclist feels, is not correlated with performance. So sometimes they feel miserable, but their performance was pretty good, and vice versa. So that's why moving from feeling-based decisions to database decision is likely to really improve the race. Well, what I love about this example is it's not man versus machine. It's man and machine together. It's like the chess player. The greatest chess player in the world is not deep blue, it's deep blue and a human, or a team of humans, actually. So where does the human leave off? I mean, you were just saying that the cyclist may feel good, but it's not necessarily an indicator of performance. So I'm wondering, David, have you prototyped this and how do you change your sort of behavior as a counterintuitive to you? Are you sort of doing things that you wouldn't normally have done? I wonder if you could talk about that. I don't think I'm doing necessarily anything different. I'm still riding my bike, and I'm watching my Garmin, and it's telling me what I'm doing. Once in a while though, I mean power, you can really see that from a riding standpoint. And I may feel tired. And in this race, there's moments when the mental parts of, so fighting some of that mental, the mental things where you're totally not with it. In fact, maybe not that coherent as to what you're doing other than riding your bike. My crew is gonna say, well, you are performing well. So that could actually just give me a boost. Or I may think I'm cruising down the road at 20 miles an hour and I'm looking at my computer and it's not quite giving me the information I want. And we may have a reason because I am tired or I'm not continuing with that power. So it's something that's gonna assist me and help me and help us make decisions because at points in the race, you think you're doing really well and you're not. Now, GFP, what about other biometrics? Chloric intake, when David's sleeping, are you measuring that? Is there that type of feedback? No, but what we're measuring, there are also some medical lactic acid rates, so we monitor also its condition. But I want to come to your point. The decisions are made by Dave and his team, not by us. So what analytics will provide them is what are the consequences? We recommend this, but we also tell them if you don't read this, but one hour later, this is the delay you will get, say five minutes. So if for five minute delay for the reason he feels good, he wants to race more, that's worth it. If we say, oh, if you delay by one hour, you will be hit by one hour, maybe they will stick to what we recommend. So it's not a one or nothing, it's really decision support. The decisions are made by him and his crew and analytics is informing the decision. So this is, again, a great example of internet of things. When did you guys start collaborating? How long has this been going on? Yeah, about a month ago. So it's a new collaboration. It's a race as well. Awesome. So we're racing to race. Fantastic, and let's see this afternoon. We're going to be broadcasting on the UltraCyclist channel at ibmvisiongo.com. So check that out. David's going to be doing this real time, working out the bugs, I presume. There's a few little things that we are working through and me learning what the information can be myself and then kind of learning each other and what we can actually do. And it's pretty amazing to me. Well, what a fantastic collaboration. Congratulations and best of luck to you, David. Thank you. And when is the race? The actual race is? The race starts June 16th. Okay, great. Hopefully it's finished in eight days. Yeah. It's our goal. Good luck. Be healthy and be data fortified, all right. Excellent. He has a job. J.F.P., thanks very much to you and David for coming on theCUBE. It was a pleasure to meet you. Good luck. Thank you. Everybody will be back with our next guest. This is theCUBE. We're live from IBM Vision in Orlando. Right back.