This week we're going to look at data science and its relationship to your analytics program and your overall work in an agile and in product development. Really, what we're after is to answer the question of as a product lead, as a product manager, how do you take what you're doing in an Agile Analytics Program and bring better inputs to your data science team and create a better collaboration with them so you can work together and get to even more awesome outcomes? Likewise, if you're a data scientist or an analyst, how do you coach your collaborators to ask better questions and go out and get the data that you want to have to help them get you guys together to a better outcome. The first question that we want to answer is, what is data science exactly? There's a Venn diagram by Drew Conway. Currently, it's a really popular way to think about the fundamentals of this relatively new now interdisciplinary area. The idea is that there is substantive expertise, and this is just domain knowledge. In the case of HVAC and Harry, for example, this would probably be the pairing of understanding the substance of the HVAC business itself, and then using the tools of design and Agile to go out and make sure that we ask the right questions at the right time. We know the life of Trent, the Technician. We know what problems, what jobs are on his or her list. We've tested our value propositions, we've iterated a good usability and we've got working software that is relevant and valuable to this user. Then we have this idea of hacking skills. What they mean by this is using the tools of coding and digital infrastructure to run more powerful tools on the data and get the data at scale that you want to have. The intersection of these two things is what's called the danger zone. The idea being that without this third area, math and statistics knowledge, you can make inferences that are wrong, basically that are inappropriate or misleading and so that's why data science is really at the intersection of all three of these things. So one question that you might get asked is, well what, and what's the difference between this and just like traditional research of, for example, we know statistics and we know a lot about fish populations. Well, we could use statistical tools to ask and answer questions about that and that's what they would call traditional research. The difference being that the hacking skills over there are a way to bring even more powerful tools to bear by using digital broadly construed to get data, get data at scale, interleave data, create relationships between data at scale that wouldn't have been possible with these other tools. You may have also heard of machine learning and at the intersection of hacking skills and math and statistics knowledge is this idea of machine learning. Like we can take a big bunch of data and stick these algorithms on it and see what interesting questions come back. One example that is, there are these unsupervised algorithms that'll just go out and look at data and show you where the big clumps of correlation are, but without substantive expertise, you're not necessarily going to be pairing those with good questions or focal objectives to get to a good place. So the idea is that data science is really at the intersection of all three of these things, and that a data scientist will be able to bring these together for you to help you ask and answer really great questions and get good answers on an ongoing basis and to make this part of your digital infrastructure and your habits, and the way that you think about your work. How might we leveraged data science here? That's what we're after this week.