Welcome to the What's Up Enterprise Podcast! On our first episode, host David Caspillo sits down with Matt Florian to learn more about his background, hear his insights on the data world, and share a few laughs! Stay tuned to meet our other two co-hosts, Robert & Barry in our upcoming episodes.
Follow Dave on Linkedin: https://www.linkedin.com/in/david-caspillo-6482aa/
Follow Matt on LinkedIn: https://www.linkedin.com/in/mattflorian/
Hey everyone, thanks for joining us. My name's David Caspillo. I'm the host of Comerit data and Analytics podcast, and today we're gonna be interviewing Matt Florian, the director of the Cloud analytics practice at com.
How's it going Matt? It's going pretty good. Dave, how are you man? I'm doing pretty good. Just want to see how these podcast things here go? We'll have to edit that out, Kylie . But anyways. Hey Matt. So can you introduce yourself and how you got into the data field? Sure. So I've been doing data.
As a my profession since it's it when I got started in college by mistake. Did you really ? I did by mistake. I got started as a work study where I worked in a a laboratory for veterinary sciences and as a work study, and I got thrown into managing a database and. Like Oracle five I had worked with, and this is pretty good for a political science student and that work progressed on, I did more work studies that involved data and eventually I'd been a political science.
I thought I was gonna go and become a attorney, but I had somebody give me an intervention, train me in Oracle and other data sciences and other data practices. And from that point on, I have been. A data practitioner. So I've been doing this since the early nineties and continue on till today and just, I have an absolute passion for data.
So let me ask you, were you going to be a political science major and go into politics first? ? I was a political science major still. That's what I graduated with, was political science later on. Got an MBA in e-commerce, but it's with an emphasis in analytics is what I came out of college with. So I was gonna be a lawyer, but like I said, there was an intervention, so I never became one.
So what makes what? What made data so exciting like the, that's actually a really good question I have for you. Why is that so exciting? How do you make a career outta data? It's interesting because I fell in love with data when I was in political science. It was being able to understand the politics of why somebody does something, why they, a policy and a person's reaction.
Response to it. And we did a lot of geographic and demographic type of analytics in college. Around political decisions, political ideologies, and that just got me really excited about data and what it could tell. And data always told a story. So when I got to use data in other areas of business and operations, the same thing happened.
Data tells a story. And no matter where I'm at, it's always the story to tell. And being able to unleash that story man is exciting. What other areas? What other data areas have you worked with? I've never been really constrained by industry, so I have worked with data back in the telco days of baby Bells.
US West and Quest that don't even exist anymore, but we did data and analytics for them on call, detail records and fun stuff like that. I've done built an oncology database for Emory University that went and studied clusters of cancers and ended up winning awards and for its innovation. What we've done.
I've done data analytics for the public sector in studying. Healthcare fraud, Medicaid fraud, and built and developed patents for Medicaid fraud. I've done it for workforce management. There's the ability to go in and apply patterns to understand data regardless of industry, is really what I've honed in on.
And you find that one industry leads to knowledge in another industry, and it's been, it's a beautiful thing quite. How do you, how does a person figure out that I would be good? As a data scientist or data practitioner? Or a data architect? What? What do you think inside of you makes you like, I lined up perfectly with this type of work.
I guess ask yourself, do you like puzzles? Do you like solving a puzzle? And that, cause that's what data is. And it's interesting, I don't like jigsaw puzzles, but I love puzzles. If I'm gonna play a game on my phone, it tends to be a puzzle. It's go and solve some mystery connect dots, a, B, and C.
But all you know are dots, A and F. It's an, it's such an interesting thing to follow and do that you're. You are a systems thinker. I think the best people I know in data think in systems they don't think very, they don't think linearly. Data in working in analytics is a very abstract concept, and if you can think in that abstract concept and think in systems, you'll find that.
data just makes sense cuz you start categorizing, you put it into a system. It's really, it's, it is a cool thing to watch. What kind of systems, when you say systems what are you referring to? Data isn't, data doesn't just appear right? Data comes to us because an event, something created this data, something.
Tell, tell me what kinds of things create data. . You walk into the front door of a store, you created data, cuz now you have entering foot traffic. That's data. You go and fill out your census form. That's data. You took an action. It's that data in itself operates in its own system to say this is data about census, but then that census data can interact with other systems.
tell me about my marketing in this particular demographic area and what, how my demographics of who my customers are matches up with the demographics from the census. Those are all data and they tell systems of tell stories of their systems. Going way back to your political science days, where were you guys getting data for your political science studies?
Census was a lot of it. The government, you, a lot of people give census data and manufacturing and sentiment analysis data back to the government. quite a bit on a regular basis. So we had that data to work with. And then there's a whole slew of survey data for analysis of sentiment analysis of political opinion analysis.
So we had that data as well, plus new world socioeconomic data, which helped us go and do analysis on policy. So today we have lots of ways in which we can collect kinds of very strange, very personal types of data. Now, on people some people call it a conspiracy, but what are your thoughts on data that's collected on you through, say, Facebook, Twitter, LinkedIn?
Do those things, collect data on people? It's their whole business model is to collect. and when you engage with a company like Facebook, LinkedIn, I'm on LinkedIn a lot, and it's, you engage knowing that your information is collected and what you share is, it's a, it's not a secret, you know it.
So you really, a person has control of what they do and say, and share on those platforms by choosing to participate and not participate. and what, how much activity you engage on you. But Yeah. But from a big corporate standpoint, like when you, you work with big corporations now is it fairly anonymous?
It's more of a collective pool of data. They're not like looking at me specifically. Or are they I don't know the underlying systems that they look at. I know that from a legality perspective, now the data is often anonymized. It's you as an individual, as a dot, and a blip of the analytics.
don't mean as much as the individual. It's a persona. Yeah. It's what personas and what other personas are like you, and that's the beautiful thing about data is that you can go and build personas of people to make certain assumptions, but they're not always correct. It doesn't, no persona can tell you the full story of who you actually are and what your opinions are gonna be, but it is a, it uses data to.
and help make a judgment. Now, these big enterprise companies, they don't need to go to a LinkedIn or a Facebook to get data. They also get data on their clients or their customers through their own mechanisms also. Are there other data collection? Maybe devices. Video cameras, how are they getting their data on their clients?
Data can come in from an any number of areas. I've worked with, I worked with a, an agricultural client not too long ago. Data comes in from tractors. Just what is being applied to the field, when is being applied? How much is going into this geospatial sector of the fields as they apply? The supply chemicals.
Supply the seeds. What, where the information come in from so many different places, and right now we have everybody say there's a flood of data coming in and there is, but we actually know what we're doing most of the time. And that flood of data, there's, you always will have outliers. People just go in harvesting information.
They know what to do with . I'm not so worried about them so much. It's they. You. They'll make guesses. Sometimes they'll be right. A clock is right twice a day. They may be right sometimes, but they're probably wrong most of the time. Matt, we're about halfway through our interview and you put the perfect segue into the question I have about what are some of your best data practices?
I think this would be important because there is so much data. Data just keeps growing in volumes. I think I reached two terabytes of personal data. , oh, that's a baby belt. You need to take more pictures. Best data practices. So I think today one of the best data practices that is out there is to start collecting and categorizing data.
That you have. It's, the interesting about it is that we don't always know when or what data connects to other data. We know data that operates within the system, the processes that created it, and it tells us and informs us a lot. But when we're able to connect data from one process and system into another, like your sales data with a, say Google Analytics behavioral data, and.
I described this to some clients as right brain, left brain activity, it gives you additional insight and you don't always know when that's gonna be. So the worst thing you could do is to not collect it. So the best practice would be collect that data, categorize it, hold it, store it, and know it's changes, and then you can actually act on it cause you can't recreate it once it's lost.
So are there actual methodologies named methodologies for data data and analytics, best practices? Are there any good books you'd recommend? So there's some really good books on just data engineering and the, there's a book that came out last year. From O'Reilly, the Data Engineering Handbook.
And I, that's been a really good seminal book on the, on that topic. And from a methodology, you hear a lot of folks talk about Data Lake and it's a, in its concept, data Lake is correct in its execution. Data lakes are usually a. because they're, they dump data into online stores like an AWS or an Azure or in Google, and then the data just accumulates, but it's never truly cataloged and become actionable.
So you data lake that is architected and intentful is significantly better than Data Lake that you just dumped data. So that kind of leads right into it. Maybe this is wrapping up. What are some of the worst practices as opposed to best practices you've seen out there? Worst practice, I'd say some of the worst practices are higher, an army of data engineers and data scientists without an actual goal of what your data platform is.
It's, you can't act on data unless you. actually understand what your parts are. It's like saying, I'm gonna go and build a car, so just give me all of the metal and pieces of rubber and I'm gonna put this thing together without a plan as to what that car looks like and you're not gonna get it. And that's what a lot of people do with data.
They dump it into a data lake and then say, all right, make something of this. It's raw material. You need to actually have some data products to work. So that you can do something consistently repeatably and find value inside of it. Okay, so I'm gonna put you on a spot. If I'm the c e o of a major corporation, retail, let's say, do I hire a data scientist or a data architect first?
Data architect first. It's why it's not even really a biased answer. It's the data architect will give you a found. Of setting up the data, that architect will go and lay out the environment that will align to your business goals of wanting data products. And it will then give a, it'll give a laboratory for a data scientist to work in.
The data, the scientist has to have a laboratory in which they're working from. They're not just out there, you trying to extract raw material, the. , they collect that raw material, and that's what the data architect's doing. It's collecting and organizing that raw material so then a scientist can come in and start exploring and finding relationships that the business may not know exists.
So the data scientists, they can, they come in and shouldn't they be around to tell me what I should be architecting? The business tells you what you should be architecting. The business tells you what the context of the data means. Data scientist doesn't know the context of the data. The data scientist is actually trying to find and discover.
Interconnectedness between processes between internal and external processes or internal external factors. The business who's just running their day-to-day and Moses going, man, I just wanna look at my reports and tell me what, how did I do today versus yesterday? That's really important. But they tell you the context cuz data.
Matters within the context in which it was created, which is, goes back to that whole system discussion. That system gave it context. The data scientist, they wanted their explorers. They look at that, but they have to still, even the best explorer had a map. That's what the data architect is. It's a map for the scientist to go and navigate and figure out where they're gonna go off the map.
To go and find something new. So it's an important point though, that the business is your key initiator really, right? They can help and work with the data architects the easiest cuz they understand the context. Is that correct? That's correct. Because they're the ones who actually created the data.
It was their process, their events. They're the ones that made it. So they have the best relationship back to what it. So that's where you actually start. So when you're out there helping clients with specifically directly, who do you work with in the the organization or the org structure when you come into these projects?
Who's your, so most of the times it hires us. Interesting. But then we go and work with the business. To be that bridge of IT and business because we will always consult to the IT organizations that it does not own this data. The generation of that data was owned by the business, and so business needs to be a equal partner, an equal collaborator in understanding and building data products that are gonna move your business to the next step.
So is that part of your consulting? Is to help really bridge the IT department and the business together in these projects to bring them together into a collaborative environment that lets them succeed? That is a lot of what we do. That's awesome. So I wanna talk to you a little bit wrapping up or towards wrapping up.
You've been a consultant now for quite some time. What do you think you'll do after you're done helping save the world through data and analytics? Oh my gosh. So my exit plan from this wild ride of data has the, I'll probably go and pick up a career in which I don't make any money. Fly fishing guide.
Quite honestly it's my passion. I see so many similarities between data and fly fishing. That moving from one to the other is a logical jump. And cause I still get to do the things I love, which is I get to teach people, I get to take them on a journey. I get to go and spend the day talking, just discussing whatever's going on and be outside.
So that is more likely what we're gonna be doing. So help me out, give me an example of how data. , a architecting data modeling and analytics. How does that integrate with fly fishing or gimme an example. How are the two related? I'm I'm very curious now. When you fly fish and I go fly fishing from Raf side fish, rivers, and I guide the.
and that's what I do as an architect, is I guide the, this journey for a data journey with clients. And as I'm guiding, I can be going down the same river that I've floated 20 times and every time I float it, it's gonna be different. There's a lot that's the same. The bends and turns of the geography are the same, but what lies underneath that river?
The rocks, the debris, things that I can't see. Those are the things that makes the. Adventurous, but at the same time, having the experience to know how to identify, how to see it early stops you from getting hung up on a rock that you didn't see. And the same thing is about data. And the same thing when you consult is the experience of being able to see the rocks ahead of you that are gonna get you caught up on your journey and stop you and become a blocker and.
Rip your boat apart if you're not careful. Those are the kind of things that I do with clients and that's why the two, you segue so nicely together because they are one and the same. So sounds like we need to take our clients out on a fly flushing trip in order to, for them to truly understand the nature of data.
I am all for that trip. , it's, I got the rods there. Sidney over there ready to go. The boat is ready to go. Just you're a little bit ice off the water and we're ready to float. So let me ask you in closing, what do you see, where are things headed in 2023? I am really optimistic for 2023. 2023 is having a lot more conversations in the data world about the resurgence of the importance of data modeling and having a data foundation, which for those of us that have been in data architecture for so long, seeing that come back and being back in vogue, it never should have been gone.
But being back in vogue and having conversations with the enterprise about the role of data modeling, How to do it pragmatically. I think that will give foundation to the enterprise on how to build a good data program and build data products across global enterprises that give you insights and your predictive and prescriptive analytics to run your business.
Last few minutes. Is there a role for artificial intelligence with all this data model? AI comes into play all the time. It's in the simplest and the most complex. Yeah. We see that manufacturing, AI using and training those models to identify defects on the line and get rid of things that the human eye can't to count faster than the human can count.
These are all really important things. I have a friend who developed a vision AI tech for the ski. to be able to identify quickly before a lift operator can see that somebody is, you clicked over and is about to have an injury. So a risk aversion ai, it, there's so many amazing things that can be done with it.
and we have the data to support it, it's only gonna get bigger. So we're gonna have to definitely bring up some of those topics in our future podcast. What other types of things can we expect to, to hear over the coming months and years in this podcast? You're gonna hear me talk a lot about data architecture, data engineer.
and the foundational pieces of being a good data manager and good data steward of the enterprise and, all that is needed in order for us to even have the conversations about ai, about machine learning and about moving the business forward from a data products. I'm a big fan of the principles of data mesh for global enterprises.
and there's your other data modeling techniques and approaches from the data that I'm strong, you're really heavy on in the automation, the balance between automation and hand coding. And there is a balance. It's most people do one all the way, one all the way to the other. And I'm a bit more pragmatic about it.
So when it comes to data, I can be a bit of a curmudgeon and not really want to jump on things that are. But use things that are pragmatic, take bits and pieces of what we know, cuz that's how you're gonna move your whole enterprise forward. So last question. If there is a name or a person out there that you would want to interview with respect to that or anything is there someone out there you're thinking about?
It would be. There's a couple people that I think really. Highlight that area about that pragmatic side on the automation piece. And it's a young company coalesce, put out by Armand Petros. , and that is their whole focus is on automating all the way up to the last mile, right?
If you think about data as part of a supply chain, you move as much as you can for the lowest cost, and then the last mile is where you spend the money on, a lot of smaller vehicles to move. And that big automation, I think, that coals has really nailed that space in there. And then I'd probably also wanna talk with some folks that have been doing innovations over in the AI
Folks, that's a wrap. Appreciate you all jumping in and joining in our interview with Matt Florian. Matt wanted to appre, wanted to thank you for coming on and sharing your experience with us, and we're looking forward to many more podcasts in the future with you. It was fun, Dave.
Look forward to talking more. All right. See ya. See ya.