 Live from Boston, Massachusetts, extracting the signal from the noise, it's theCUBE, covering HP Big Data Conference 2015, brought to you by HP Software. Now, your hosts, John Furrier and Dave Vellante. Back to day two coverage of Silicon Angles theCUBE. This is our flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier with Silicon Angle, my co-host Dave Vellante with Wikibon.com and our next guest is Zach Matthew, Director of Analytics at HMatrix. Welcome to theCUBE. Hi. So tell us a little bit about what you guys do and why you're here at HP Big Data. So I work as an analytics director in HMatrix and what HMatrix does is we're a business intelligence and data analytics solution provider in the healthcare space. When we looked at the problem six years ago, what we found is traditional reporting happening. The end user, the people who do the real care. The real work. The real work, struggling to get data to base their decisions on. And so what we did is we provided easy to manage or easy to understand, point and click the user-friendly environment for non-technical users and so that they could make their decisions quickly and do what they do best. That means taking care of their patients. So talk about the role of data analysts versus data science. Yesterday in the keynote, like Stonebreaker was saying, it's hard to be a data analyst than move to data science unless you have a statistics background. And people are really looking at this new role because you have on one camp, data analysts who are doing some great work, you have developers building solutions for non-technical users like yourself, your company, and then you have the hardcore data science. You guys writing code, doing some stuff. Where is this, how do you make sense of this in your mind? How do you talk to your friends and colleagues around this evolution? I think data science is here to stay. There is, is there a value just being a data scientist? I'm not really sure about it, but there is a lot of value if we can combine data science with the business acumen. So let's say for example, if you are in banking or in healthcare, if you know what you're after, if you know the business problem, definitely applying data science principles to it will help you honor the goal or the key or solve the business problem. So that's the tremendous value out of it. Data analysis, yes, it will help honor the insights from what you have. So the big data platform has now allowed us to store a lot of data and mine it as you need it. But at the end, what you need is combine the business acumen with the data science principles to unearth what, to solve your business problems. How about the data acquisition piece, right? One of the things that came up yesterday was data variety. And it's always hard to get data sources aligned with databases and schemas and unstructured. And how is that creating innovation? Because you mentioned a good point there. The more you have on the data science and you might have an outlier that explodes into a great insight. That might not have been captured before. Take us through that kind of mindset workflow. How does someone tackle? How does someone get to that outcome? Is there a path? Is there a best practice? So one of the things that Nate referred to today, Nate referred to today was make more errors, find more and make less errors in the process. And then later on, there was mentioned that big data doesn't mean that there's going to be more accurate or there's going to be more questions, more data available. Hence, there'll be more insights that you'll get to. But the initial insights, how you translate it, how you translate it into solving your business problems is kind of key in this particular case. You may have, so the principle what we have followed is as follows, define our business problem, then identify the people who can solve your problem. It's easier to buy technology nowadays than finding the right people. Finding the technology comes last. And in many cases, what they've seen is the reverse. People invest in technology first and then try to figure out the solution. And in my opinion, that's the wrong approach. Okay, we were talking earlier on the kickoff and yesterday, people process technology in that order. We try to force some technology on the people that aren't ready for it. You've probably seen that example. Can you share a story and how do you guys, how do you manage that process? Because it's easy to just jump out of solution from a sales guy or a new shiny technology. Yeah, so there was one example that we had a couple of months ago. So one of the senior executives was asking, oh, we're going to buy this Hadoop infrastructure. And then the first question I asked was, why? And there was a pause. And at that point, it was very evident that- What do you mean? Do I need to explain? Do I need to explain? Why did you do that? You didn't have your business problem lined out. So whenever we try to be technology agnostic, when I say technology agnostic, big data is great, there are these additional capabilities. But unless you start with a business problem and identify the key questions that you want to answer, there could be other insights that come along with it. But unless you identify the first big questions, I think you're going in the wrong direction. So Ken Rudin made the point that you made earlier, which is it's not just about the insights, it's about being able to actually use those insights to affect change in the business and have a positive business outcome. You talked before about getting information and insights into the hands of non-technical users, presumably so they could act on it. That's kind of the holy grail of big data. Sort of the citizen data scientist, if you will. So I wonder if you could paint a picture of specifically what you're doing in that regard. What the business problem is, how you're approaching it, the technology behind it, and the people in process pieces. Paint a picture for us, exactly. So let me give you one example of how we use big data and then we'll get into one particular application. So with using Vertica, powered by Vertica, what we saw as immediate gains was threefold. One, the improved response type. That's a no-brainer. When we do our comparison, what we found is like a 70 times improved response type. That is great. That's number one. Earlier, preparing the data was kind of done by a siloed information technology or some other group. In this particular case, because the governance and administration out of this is very simple, that preparation time or the preparation constraint from a DBA or from a systems person is almost negligent. So you eliminated one layer of bureaucracy right away. And then the preparation. So once, let's say, the information is loaded. Let's say, how many patients do you have? What are my patients doing now? For a nurse or somebody who's non-technical to look at it, earlier they would have to wait for a report generator who would take their own sweet time. And that additional bureaucracy has been eliminated because of this point in time. It's collapsed at a lapse time. Yes. Between I need an answer and getting an answer. Right, so there are three levels of improvements that you've done just by bypassing, by using the improved response time, eliminated the systems on the front end and on the back end. So you've eliminated the systems in the front end. Maybe you could give us a historical view. Or do you have an enterprise data warehouse and you're doing ETL offloads from that? Putting them into Vertica or is Vertica your EDW? Maybe you could describe for that whole time. Yeah, so at this point, we're looking at claims. Because from a hard of healthcare claims right? So even though, when you say claims as general, no two providers are going to have the same set of structural claims. Everything's going to be different. Within a broader claims, the subject is the same. But once we get the claims from the client, regardless of whether it's a provider or a payer, we load it into our Vertica warehouse. There is business intelligence visualization on top of it, which is non-technical, user-friendly point-and-click which nurses or any businesses can use. Now, one of the things, or one of the applications that we are very proud of is like a population health management tool. So for care managers who are kind of managing a lot of patients, they can get overall numbers of how many patients are there. They can implement programs, they can get feedback on how they are performing. And then for visiting nurses who really take care of the patients, they can go to remote locations or they can sit in their office, they can enter their patient information or they can monitor it or they can have customized alerts. Let's say if you make certain alerts when your blood pressure goes down or diabetes levels change. So at the end, what we do is eliminate the bureaucracy so that care managers do what they do best, taking care of their patients. We have a question from the crowd here on the crowdchat.net slash HP Big Data 2015. How effective is agile for infrastructure and big data projects in your opinion within the healthcare space? Is agile a known term? And then second question I want to add to that for me, for me is what research and cutting edge stuff are you looking at? Because when you look at agile, you're looking at the new way of architecting these kinds of solutions. So is agile effective for infrastructure and big data projects in healthcare? And two, what are you personally looking at with the Cornell, which has got a lot of big data stuff going on there, as well as other universities? Where's the cutting edge action going on for new technology research? So in general, agile I think is there to stay. And one of the reasons why I say that is being a small company, we are nimble and agile. And the typical, I used to work for a big financial institution. I know the typical- Waterfall. I know the typical project lifecycle timelines. But in this particular case, by the time the planning phase of a typical, traditional project gets over, our execution time or delivery is over. So we continuously, all our customers, I would say, give that customer satisfaction because we are nimble and agile and get things done. So you think that's a good direction? That's a good direction. And with the transition to cloud with cloud technology emerging and being so cheap, it is very much there in the architecture side too. Another question is, are you working with payers, insurers, or providers? So our clients include employers, payers, providers, benefit consultants, public sector, and people who depend on them. Okay, so research, what areas are you looking at? Just on a personal level, in the big data, a lot of stuff's happening, a lot of new stuff's being developed. What are you looking at for research? Okay, so there are analytics models that power our business intelligence environment. So those, the improvements in modeling, the improvements in statistics, improvements in analytics is always what is there in our mind. How did you get into this whole big data thing? What are you in? That's an interesting question. So I used to work for one of the big financial firms and the typical question that when I go to like a big project meeting or a big board meeting was such that, where do we place a new bank branch? And the typical take away from one meeting would be, okay, let's figure out, you know, let's go to this department, get the answer for it, and then come back and have another meeting. By the time the next meeting comes up, it would be two weeks down the line. And by the time, meeting after meeting, you would end up having six months of meeting without no result. And then I go to this, you know, this company Hmetrics, you know, the current owner is my former boss. So he called me and asked, you know, let's see what we do. So I come over there and then I look at certain patient information and I ask, how did you, what is it based on? He said, oh, it's live database, everything is real time. I said, how many records are here? It's like a couple of billions. And then he said, there is something wrong here. I could not believe it because at that time, it was five years ago, I didn't know about Big Gear technology. And I was not, I thought that there was something he was tricking me on. And it turned out that it was not and I got all excited. The term was not Big Gear at that time, but that was my evangelism into Big Data technology. Zach, I really appreciate you coming on theCUBE. Final question is, what do you think about this event? This event here, HP Big Data, the makeup, is it like other events, is it unique? What's your take on the vibe and the people here? The fact that we were able to meet Mr. Stonebreaker yesterday itself made it for it. What I think about this event is this is less, I would say, fluff or less marketing and more meat. So I'm very glad I attended it. I hope to have a nice time further in the next two days too. As we say, meet on the bone versus the sizzle. Sizzle on the steak, more steak, let's sizzle. This is theCUBE, of course, bringing you all the data. We'll be right back here from live from Boston after this short break.