dave-kelly-mixed.mp3 [00:00:00] If you're a data scientist, I think there's something even more meaningful you can do. So I really would encourage people to to get involved with nonprofits, offer your skills, because I've as I've said earlier, you know, I just think there's so much information that can be can be gleaned from Data. And sometimes they're just great opportunities to learn from Data and do things a little bit better. [00:00:42] What's up, everybody? Welcome to the artists Data Science podcast, the only self development podcast for Data scientists. You're going to learn from and be inspired by the people, ideas and conversations that'll encourage creativity and innovation in yourself so that you can do the same for others. I also hosted Open Office Hours. You can register to attend by going to Italy dot com forward. Slash a d. S o h. I look forward to seeing you all there. Let's ride this beat out into another awesome episode. And don't forget to subscribe to the show and leave a five star review. Our guest today is a data analytics industry veteran and entrepreneur. He's an alumnus from the University of North Florida and has earned his MBA from Georgia Tech. [00:01:42] He started his career in 1995 at Equifax and has since gone on to start two successful companies, Sigma Analytics in 1997, which is acquired by Merkle and Analytics IQ in 2006. Analytics IQ is a dynamic, fast growing marketing Data and predictive analytics company that is focused on providing innovative consumer data and analytics solutions. As CEO, he puts culture at the forefront of his company and is proud to work with the tight knit group of Data fanatics who love to move quickly and make things happen for their clients. He's also the founder of a nonprofit organization called Honey and Haven, with the mission of supporting an orphanage located in Aganga, Uganda, that helps over 80 children and 11 local women by providing them with work in the orphanage. So please help me in welcoming our guest today, CEO of Analytics IQ, Dave Kelly. Dave, thank you so much for taking time at your schedule to come on to the show today. I really appreciate you being here. Yeah, thanks, Harp. I'm happy to be here. And, man, I hope I did not butcher the name of that. Did did I say that right? I want to do do this town. Justice is going gonna be gone baby gone. That's awesome. And I'd love to hear about how you got how you got involved in that. But before we do that, let's learn a little bit more about you. So where did you grow up and what was it like there? [00:03:07] Ok, I grew up in Jacksonville, Florida. It was hot. You know, the thing it really had going for it was the ocean. So I grew up surfing. You know, I spent a lot of time on the beach. But as a kid, I always believe that all roads at that time led to Atlanta, probably literally and figuratively. And I always wanted to move to Atlanta. So when I had the opportunity to move or when I decided to go to graduate school, I had to pick Georgia Tech. And I've been here for nearly 30 years now. [00:03:41] So is is Atlanta kind of like the metropolis of that region where you're from? [00:03:46] It really is. You know, at least when I was a kid, there really wasn't much going on in Jacksonville. And, you know, Atlanta was where all the sports teams were. It's where the business activity was and still is. So, you know, I always looked at Atlanta, so, wow, it'd be great to live there. And then at some point I did. [00:04:06] So I'm I'm originally from from California. So surfer culture is very, very much ingrained in me. Even though I don't know how to surf. I've never surfed. I just go to the beach and love watching them surf. So I didn't I didn't know that Jacksonville had a surfing culture in the surfing community. They're really pretty heavily involved in that. You surf a lot or at all? [00:04:25] Well, I did surf quite a bit. And, you know, our waves aren't great compared to California, but we had a little something. So you actually could do it. You know, it was fun. I haven't actually tried it in a while. It would be interesting to see if it's like riding a bike, if I could just hop on a surfboard and and do it again. But it was something I used to enjoy. But I still love the ocean. I love anything connected to the water. And if I had one major criticism of Atlanta besides the traffic, it would be lack of a body of water. [00:04:58] I'm very, very much the same way, like, oh, I love I love bodies of water. Like, I love my my dream in life is to just have this basement office that I'm sitting and just be something that overlooks a nice lake with some mountains or something. I just love being next to water. It's something very peaceful and tranquil about that, isn't it? Totally agree. Yeah, I totally agree. So that's awesome background hearing about growing up in Jacksonville. But, you know, when you're in high school, when you're right around that age, what did you think your future would look like? Did you think that it would be something that analytics would be something that you'd be involved in? I didn't know. I hadn't I had no idea. [00:05:36] I've always been really good at math, probably like a lot of your listeners. But I honestly didn't even know it was a profession. You know, like a lot of people in life, I sort of had meandered along and come to a fork in the road and you sort of decide which way you're going to go. And I just say that's try the best description of how I've ended up where I am. There wasn't because, you know, I was Steve Jobs or something and had some vision, you know, when I was really little, it sort of took advantage of opportunities as they came up along the way. [00:06:08] So what was it that you wanted to be when you're in high school? What did you think you're going to be when you grow up? [00:06:13] An astronomer? Yeah, that was and I'm still fascinated by astronomy and astrophysics, but that was really what I wanted to do. Why didn't. I'm not sure. But I would say from age like four to age sixteen. [00:06:32] That was my obsession when I was in early high school. Very much so obsessed with them, with astronomy as well. I remember back in those days, I think it was there, we had a 486 computer that just got connected to the Internet. So I'm pretty old. I'm tired old school here. I would spend a large majority of my time surfing the web and just researching about black holes and and the galaxy and universe and things like that. Just countless hours doing that stuff I'm so fascinated by. [00:06:59] And I read a lot about it, but it's just a hobby now. [00:07:03] Any any particular books you read about it or is there any source that you go to? [00:07:08] Well, you know, right now I'm reading a great book by Avi Loeb. He's he heads up the astronomy department at Harvard. And he just came out about Momoa. If you're if you recall that thing from outside of our solar system that kind of came shooting through a couple of years ago. Yeah, he's he believes that it's artificially constructed and he has a lot of really interesting Data in this book. So like I said, I find it totally fascinating. Check that out. [00:07:42] So, OK, so you went from from wanting to be an astronomer into astronomy and then into a database marketing. So. So what what is what is database marketing and how did you become interested in it? [00:07:52] Well, you know, Data is marketing to me, from my perspective, is really applying science to Data to market smarter. And, you know, like like I said earlier, I just sort of meandered my way into that space. I had, you know, one thing about me, I've always wanted to be an entrepreneur, and my first startup was in the risk analytic space and back and what I call the good old days when you could make a living building like credit scoring models and risk models for large companies. And I sold the company to a large marketing agency and for the first time got exposed to marketing Data and at the time, and this time only being 13, 14 years ago, there wasn't a lot of science applied to marketing Data you know, a lot of people still call it comp. Data meaning implying that it's more of an IT function than an analytics function. So when I started analytics IQ, the focus was totally on taking a scientific approach to creating marketing data and predicting smarter. [00:09:03] Actually, never heard that that terminology before at compiled data. That's a new one for me. What exactly does that mean? [00:09:10] Well, I think it refers to if you look way back in twenty years ago, that the way marketing data was created, it was little bits and pieces of information that you compiled together. So instead of saying, hey, let's take an analytical approach, let's model this, because that might be a better way than just sticking it together. I think that the compiled name is still stuck around referring to the way that kind of data used to be put together. [00:09:44] How did you how did you kind of start analytics IQ? Like what was the the opportunity you saw in the market and how did that spark this idea for starting this company? [00:09:56] Well, I'm a big believer that if you if you or any of your listeners ever jump out and start your own company, you need to be really sure that you have a competitive advantage of some sort. Your advantage might just be that you have a pricing advantage. But to me, it's even better to say, well, I've got a better mousetrap in the world. Who cares about better mouse traps? In this case, I had a front row seat to one of the really large marketing agencies. Kind of got to see firsthand how I would say inaccurate a lot of the marketing Data was. And I just believe there was an opportunity to do it better. So I jumped out and started a new company and started acquiring data sets and hiring data scientists, and we started from scratch building our own marketing tools. [00:10:51] So what would you say, analytics IQ like the what's the company all about? Like what's what would you say the mission statement of the company is? [00:10:59] So our mission is to create more relevant information about humans with the intent that with this information, marketers can put much more relevant information in front of you. You know, for example, you know, I my youngest child is 18, so I really need to sit through diaper commercials, for example, on one hand. But on the other hand, do I need commercials aimed at retirement? So with. I believe. With better Data, you can create a more relevant experience for people, and I've known from the beginning of the Analects IQ to use, you know, another analogy, we build hammers, we don't build houses, but we build a tool that we think can be really relevant to building a better house. And our mission is just to build the best Tamar's possible. So like I said from the very beginning, everything we've we've ever built, we validated. We make sure that, you know, independently that, hey, we're predicting that this household earns a cat. Well, we validate that it's a prediction. It's not a fact. But we our mission is to do something that's much, much better than knowing nothing about the household. [00:12:13] It's really interesting. And just like just kind of fascinating how you able to to create these datasets, I guess, without sharing too much of the secret sauce here. Like, how does that work? Like how does this process work to create these data sets that you create? I mean, we'll talk about one of your big ones a little bit later. But is it just a matter of I'm going to call up every house in the neighborhood, hey, do you own cat? [00:12:36] So. So we do surveys. And the general idea is you can know something about a small group of people and then you can project it and predict it on a larger group of people. A lot of your listeners, I'm sure, familiar with the concept and data science. So we do a lot of surveys. So to use the cat example, perhaps we reach out to ten thousand individuals and ask them a pretty simple question. Do you own a cat or not? And I think that's easy because everybody knows whether they own a cat and people have no motivation to lie about it. So and that's important. And we can talk about later why I say that. But certain things are pretty straightforward in this case. Let's say that out of ten thousand people, two thousand people say, yes, I own a cat. Eight thousand people say, no, I don't own a cat. Well, so we we wade into the data and say, OK, what's different about people that own cats versus people that don't? And I don't have it in front of me. But let's just say that people that are single or more likely to own cats, perhaps maybe people that live closer to a city center so they're less likely to have a large yard, maybe a certain age range, maybe skews towards a gender. [00:13:53] So we build a model and we say, OK, we we like it. We were able to predict this with this all this independent data and we'll project it out to our entire universe. So in our case, at the moment, we're focused on the USA. So we're we're projected out to the universe of two hundred million US adults. Well, independently validated with another survey. But when it's all done, we'll go. OK, we now have a universe of twenty million individuals. Twenty million households. I mean, where we we believe they own a cat and are independent validation is said there were eighty percent accurate that eighty percent of these people do indeed own cats. So if you're, you know, a large if you're a company that markets products to cat owners like cat litter or cat food, you in turn will pass a royalty for using this this model that we created. So that's a fundamentally our business. You know, on a good month, we collect royalties on around a billion transactions. So digital ads that are displayed or there's a decent chance it's one of our models is behind it. [00:14:56] That is really, really fascinating. That's super cool, I guess. But that he had and it definitely really interested in learning more about that. And we'll get to that here shortly. Thank you for giving us an idea of what analytics IQ does and kind of what the mission statements are about and how you guys put together these data sets. But what's the change that you want to see in the world as a result of starting this company, especially someone who's involved in a nonprofit? [00:15:20] Also, I would say I can't claim that, you know, I'm solving global warming or something with the core business that the company has. But I would say I do think and not everybody agrees with this, but I do think the world obviously there a slightly better place. If you have a more relevant experience when you're online, when you're watching smart television in terms of, OK, well, at least relevant products are being put in front of me. The other thing I think is, you know, right at the moment we have thirty five employees and my hope is we're providing a better life for these employees and their families. So, you know, I look at both those things and I feel good about it. [00:16:09] That's pretty interesting because back in the days when we didn't have smart TVs, we just got whatever commercials were put in front of us. Right. And so nowadays it's you know, if you're watching stuff on YouTube or what have you, the ads are more tailored towards what it is that you're interested in. So kind of in the background of what's happening, it could be some of the work that. And looks like he was doing to help us some of this relevant stuff, right? [00:16:34] It very well could be. And, you know, smart TV is kind of in the early stages. But more and more, you know, if you watch things on demand, especially watch things in demand with commercials, you know, again, the the provider has the opportunity to use Data to target the commercials. I don't think it always happens, but increasingly it does. [00:16:58] Really fascinating. And I mean, this is all kind of like a new world to me in the sense that this is how we can leverage these compile data sets to to help make better experiences for for people in terms of, you know, whether it's putting relevant products in front of them or giving them content. How else can you see Data scientists working inside of larger companies benefit from using external data sources? [00:17:23] Well, no matter how much good first party data you have, first party referring to typically known data about your customers, it's it's often helpful to get an outside perspective and to know more. And increasingly, we're trying to focus on things you can't know about people. You could only predict them. For example, perhaps you can know my age. You can know by gender. You can know what kind of car I drive. Those are all things that are possible to know. You can't know how important the environment is to me. You can only predict it, right. So increasingly, we're focused more on things that you can only predict about people. And, you know, we we believe we have a place at the table. And in many, many cases, even when these credit card companies have really rich data, they still find that we can add some value into them, understanding their customers better and messaging their customers better. [00:18:23] And that's really interesting because it's adding like a whole nother dimension to the data. And this is something that Nathalie's IQ does kind of uniquely, and it's the cognitive psychology type of dimension. So talk to us about this. How is it that you guys are incorporating this type of data into data set and then maybe talk to us about the people core data set? [00:18:47] Sure. So we have several cognitive psychologists that work for us. And, you know, some years ago we started to well, actually, when we first started in surveys, we started we understood that sometimes people don't know things about themselves. Sometimes people, I don't think maliciously lied, but they do lie about sensitive subjects, maybe even to themselves. And we realized that we needed to take a different approach to collect the kind of data that we wanted. So thus, you know, first partnering with the cognitive psychologists and then hiring our own cognitive psychologists. We treat it like a personality test you. It's really difficult to come out and ask someone their motivation for donating. You can I just don't think people always know what it is. So we treat it like a personality survey where you ask questions different ways. You you ask pieces of a question and you stick the question together. At the end, you throw out deceptive answers. And again, it is very similar to how personality tests are done. [00:20:01] So it's kind of like the like for example, speaking of personality types, like the Mires break test, for example, that's the popular I think a lot of people, a lot of people taking that. [00:20:10] And if you have again, it just it doesn't ask you three questions. It asks you a number of questions to 2s, slot you in to a personality type. And I would say that's the exact same approach we take with things we do. There might be a relatively simple thing that we're modeling, like whether someone has extrinsic motivations for donating to a nonprofit, meaning they want the attention, they want their name on a building. The opposite of that is intrinsic meaning. I don't want anybody to know that I'm involved with. I'm donating. That's a spectrum. But to get to that, we we ask a lot of different types of questions so that we feel confident that we've correctly stratified people based on the particular thing that we're trying to predict before we model it out to our universe. So, you know, we we deal with PhD cognitive psychologists who've made it their life's mission to do what we're trying to do. And honestly, it was difficult at first to pull these things together because they don't naturally fit together. But it's kind of been awesome. And I feel like we've been able to create things that that we can see really work. And they're not the run of the mill. You know, again, just trying to predict someone's income just a little bit. Better than our competitors, it's trying to go into unchartered territory and do something completely unique, which, you know, we as Data scientists find exciting and I found it interesting, like they like, you know, you're talking about people might have a motivation to lie, right? [00:21:44] People don't necessarily have a motivation to lie about the the number of cats they have or if they have a candidate, but they might be motivated to lie about some other things. How is it that you structure your questions to kind of get past that? And I hope hopefully not taking too much into secret sauce here. Let me know. But I just find that really fascinating. [00:22:03] I try to think of a good example. So, for example, we have a model that predicts whether someone or to the degree that an individual suffers from stress. And the reason that's valuable is, you know, certain pharmaceutical companies. It's interesting information for them. And again, it's something you could only predict. And it's really not just as simple as saying, hey, do you suffer from stress? You know, put yes or no. We usually ask corollaries like, you know, how well do you sleep? Well, I feel like that's something people can usually answer correctly. You know, we ask how many hours a week someone works. We look at these corollaries and then at the end you pull it together and say, OK, this is a group that we're confident suffers from stress. And this is a group that we're confident that they're the opposite. They have little or no stress in their life. So, again, you know, it's it's dancing around something, especially a sensitive subject to get to what you're really trying to understand. [00:23:09] And just imagine how much fun the data scientist and the data analyst that work at your company have with this type of data, because there's that's just opens up so many different ways to do exploratory data analysis and things like that. Super, super fascinating. You guys are hiring. Let me know. And we're growing. And we are hiring. Yes. They will definitely be sure to drop a link in the in the show notes that people can apply for for that. That's really interesting. So for this, people called Data said so is this like the domain Data that you have people call in? Are there other data sets that you have? You know what's kind of unique about the people? Chordata Data. [00:23:48] Ok, so we have two data sets. We have people core, which is focuses on US consumers, and then we have a business core which focuses on US companies and companies is sort of an aside. We take essentially the same approach of like, OK, you know, what can you what would be valuable that you can only predict about people within companies? Because humans are still behind companies and humans are the ones making decisions. So cognitive psychology, you know, trying to predict things over there. People core is all about us consumers. And I believe we have around fifteen hundred things that we've predicted and we add fifty to one hundred per quarter. So we're constantly doing more surveys, building out new tools, validating tools, occasionally throwing tools out. I will say there have been things in the past we've tried to do and what we never were successfully able to. But if but many, many things we think you can predict with our type of data so that people cause our term for all of these attributes as a group. [00:24:52] And on the consumer side, how big is this data set or is it like billions of columns and millions of like, how massive is this? [00:25:00] Well, you know, some talking to data scientists. I mean, you know, it's two hundred two hundred million rows. So there is a row for each adult U.S. consumer. So for me, we do. Dave Kelly, one, two, three, Main Street. But then here are all these columns. It's these fifteen hundred things we predict about David Kelly, his email. Here's his age, you know, these different things that we predict. That's kind of what it looks like. The reality in the modern marketing world is, you know, while it sits or where humans live, humans don't physically live online. They live in homes that are geographically somewhere. We converted to IP addresses and mobile IDs and other ways that we can connect to consumers. [00:25:51] I feel like nowadays I'm living more and more on the Web and look at what trends have you noticed with respect to the need for data sets like this in this pandemic age that we are currently going through? [00:26:04] Well, you know, when this thing first started, I would have said, well, I have no idea what's going to happen, just like everyone else. But as it turns out, screen time has gone up a lot, like tremendously. And be honest with you, it's while the pandemic has been a tragedy, no, definitely don't want to minimize that. It has been good for our company in the sense that demand has gone up. So I. Never would have dreamed that we had a a stay at home kind of a company that benefited from the modern economy, but as it turns out, we've been able to pivot and it has parts of our business are down. We used to work with cruise lines and other industries that are way off, but the parts that are up more have more than made up for it. [00:26:52] And I found it really interesting that you guys were able to identify this, that kind of feature or persona, rather pandemic persona. So talk to us about about that, maybe share how people's attitudes are intersecting with the way they act during these times. [00:27:09] I'm sure everybody listening to this is fully aware, no matter where they are in the world. You know, there's a spectrum of concern about safety, I believe, from people I know who seldom leave their homes. They don't take their masks off. They certainly wouldn't eat in a restaurant or anything that at least right now to other people who seem to have no concern whatsoever. And, you know, they hate to wear a mask. They only do it if they have to. So it's a spectrum. And we set out to and we believe that we could classify people based on that spectrum. And we have been able to actually so we do segment individuals based on where they are on the safety spectrum in some other personas around the pandemic. From our perspective, it's something that we can sort of showcase what we can do. We always like to step in to new trends because, you know, we if you look at our business model, one thing that I think as a business person, as special as we can take an idea and turn it into a validated Data product like extremely quickly. So with this, we knew that our competitors weren't going to be able to put anything out about the pandemic and and they still won't be able to for quite a while because of the nature of the way Data is created. But it's something we could do quickly. So that is something we did. I believe we came out with it last summer. So something we were able to have an idea and get it out there. And, you know, we we make it available for some worthy causes with course, without us taking a fee. We made it available to some researchers who are trying to understand how better to react to the pandemic. And, you know, certainly we most likely have made some money on it, too, from other sources. [00:29:05] I think when one question it kind of naturally pops into my head is how do we make sure that Data is managed in a safe way so that we're protecting individual people's Data, people's identity, things like that. [00:29:20] But it's a very, very sensitive subject. And if you look back, I guess it would have been four years ago now, five years ago now in Cambridge, Analytica was in the news quite a bit, at least here in the States. And, you know, for in my opinion, misusing private data, we don't use any private data. So when I say we're building a model, it literally is taking things like census data and public record data. That's by individuals doesn't mean a lot, but that's where data science enters. So we have around a hundred different data sets that we use. None of them involved involve, you know, any kind of patient data. We don't use any kind of scraped Internet data. We don't use tracking data. So we just our model is do the best we can with data that's totally compliant with the way the world is moving. You know, this is a kind of an odd thing to say, but regulation doesn't bother me. I think it actually bothers our competitors so much that it's good for us. So and there is increasing regulation. CCPOA is a California regulation around. Privacy, again, is good for us because these are things we've always done. So it is a concern we have. We don't also we don't keep the individual research. You know, when we do surveys, we don't keep the individual replies either. So even though personne may have responded to a survey, we technically could know something about them. We throw that out. So when people core is not where they told us, that's what we predict. [00:30:57] That's really cool. That's awesome. You guys are still ahead of the curve with respect to privacy and all that stuff. It's really not an area that I've had to focus on in my most recent job. But previously it was definitely a big concern. I think a lot of up and coming data scientist and data analyst could do well by researching some of these types of requirements, like you mentioned, CCPOA, GDP and things like that. Thanks. Thanks for talking about that. You mentioned that you've made data that the data set available to nonprofits and and essentially making available to do good with. So how can. Can we use that data set like people call for good? [00:31:32] Well, we think, you know, knowing more about people is always helpful. You know, we we do work with nonprofits. We work with some large ones who are clients. So there's a financial arrangement. But and we work with one of the largest children's cancer charities in the world. And, you know, we were passionate about what we do with them. We you know, while it's always good to see your clients do well, we really, really want them to do well. And it's something that we're all passionate about. And you know what? We've we have found that better Data helps these nonprofits get better marketing results and decreases their marketing costs. So we feel like we are helping a bit. And, you know, we're we're always open to working with there are nonprofits where they're not reasonable for them to buy our Data. And we're always happy to work with causes that are that we deem worthy to help them do better. And if they can't afford to pay us, they can afford to pay us. That's OK. [00:32:43] Yeah, that's really cool, man. So, I mean, we have a lot of listeners who are up and coming Data scientist and students. And plus their science itself is just a type of field where you can pick up projects, you can pick up Data and you can start start doing things and start making a change almost immediately. So how can somebody who's armed with nothing but a laptop and then an Internet connection use data and analytics for good? [00:33:10] Well, there's a couple thoughts there. And one thing I did want to mention, we also make our data available to universities so anybody is listening and it'd have to be us because that's where our data is. But we at no charge make our data available to universities. We want data science students to play with our data to get used to it. Just like when I was in grad school, you know, I got exposed to using SAS and, you know, I don't know how much, you know, Georgia Tech paid SAS, but at a minimum, you come out of school going, OK, this is what I know how to use. So that's what I use in the future. So we want people to with our Data, we make it available to universities again, if anybody's listening and they like our Data be available at their university, certainly reach out to us. And again, no charge. We just want people to to use our Data. But more specifically to your question, in my opinion, there's just a ton of Data waiting to be analyzed. There are so many insights that could be had from Data. So I would definitely encourage everyone to volunteer for their their particular skill. I will say, having got involved with this orphanage and the orphanage had been around for quite a while, you know, bringing the special insight that I have around Data science there, there were things that you can learn to be more efficient. And again, one thing that all enterprises have in common these days is Data. And I like I said, I still believe there's just tons of Data waiting to be analyzed and for, like, really meaningful insights to come from that analysis. [00:34:51] Thank you so much for for for sharing that. I'm definitely going to include a link to the world. Maybe I'll get a link from you, rather include then to the show notes where people can reach out if they're part of a university to get some some of your data to help with whatever research they're doing. We're also going to give you guys have access to a small little sample of the people who are data set as well. I'll include a link to that in the show. Notes maybe help provide our listeners with some ideas or maybe some some tips for maybe some small initiatives that they can take up on their own for leveraging Data for good. [00:35:32] Well, like I said, I have one regret I have is I actually, in my opinion, did not do enough that the world I mean, I was raising kids and starting businesses, but that's not a great excuse. I would encourage everyone to volunteer their time. And whereas some people, you know, what they can do for a nonprofit is, you know, stand at a table at a five K race or something to help. But if you're a data scientist, I think there's something even more meaningful you can do. So I really would encourage people to to get involved with nonprofits, you know, offer your skills, because I've as I've said earlier, you know, I just think there's so much information that can be can be gleaned from Data. And, you know, one thing about nonprofits is and this is sort of stereotype. They're not always run by business people, and sometimes they're just great opportunities to learn from Data and do things a little bit better. [00:36:37] Do you have any examples of of of one of your partnerships that you've done with a nonprofit that leverage your data for their cause? And how was it that you guys worked together? What was the problem that they were able to to help solve? [00:36:52] Sure. So as I mentioned earlier, you know, we work with one of the largest children's cancer charities, at least in the US. And with what they do, I'd say the problem that we're trying to solve is to getting better marketing results. One way that we attack that was especially when we started building these personas and digging into people a little bit more was what kind of message really resonates with individuals. So this particular charity, they've done great things in curing or moving towards a cure for childhood leukemia. The childhood survival rate has gone up tremendously since this charity has been around. But and that's that's great. That's a fact that would appeal to anybody. And let's just hypothetically say they've increased the survival rate by 50 percent. On another note, there's the what I would call the tear-jerker side of a charity where perhaps it's a picture of a child who's clearly suffering from cancer. So what we found is that while the fact that the child goes up appeals to everybody, the picture of the child suffering from cancer appeals to everybody, you know, in terms of motivating some individuals respond better to one over the other. [00:38:20] So one thing that we did was dig in and say, can we divide people based on how let's just say conscientious. They are meaning, conscientious, meaning. They want a lot of facts before they make a decision. Versus the other end of that spectrum is impulsive, meaning they make people who make decisions based on emotion. They make immediate decisions. So one thing that we've successfully done is say, OK, let's divide people based on this spectrum so that we can put a message in front of people that are conscientious of fact and that appeals to these people. It's it's information. It's Data. They process it. They make a donation. On the other hand, individuals who respond better to an emotional prompt. Well, great. We put this picture that is motivating and it gets these individuals also to make a donation. So, you know, it's a nuance. You know, it's it's something that perhaps gets 10 percent better results. But when you start talking about large numbers and a really large marketing budget, that that actually makes a difference. [00:39:25] Yeah, absolutely. Especially when you're talking about that at scale, 10 percent can make a huge difference. So that's really cool. So I guess anybody listening now that you kind of have an idea of how you can leverage external data to help a nonprofit organization, maybe something that you can do for your local community is maybe find a homeless shelter in your local community or reach out to them. Right now that, you know, analytics HQ exists, maybe you guys can come into a partnership with analytics like you get some get some data and combine it with whatever their donors information they have and maybe help them design a campaign to get more donations to help with essentially helping the homeless in your community. Would that be kind of a one way we can use that? It's a great idea. It's totally a great idea. So do you have a project that, you know, if you had time, you think would be pretty interesting, pretty cool to do with one of these free data sets that you're going to be giving our audience? [00:40:25] Well, you know, I think that it's really important for newer data scientists to play with data to understand the difference between things that cause something versus things that are just a correlation of something. [00:40:43] So, you know, example always is because I'm super tall that I always say, well, and I'm six or five and I wear a size 13 shoes. So does that mean that because if you look at the inverse of tall people, maybe almost all of them have big feet, does that mean having big feet makes you tall, or is it just something that is a side effect of being tall? And I found that on their first day with us, a lot of data scientists coming from out of grad school is something that has to be learned, like, OK, maybe you can't get all the answers just from the Data. Maybe there are things you need to think about or it's important to know you have more information than just put some data in front of me and let me start crunching numbers so I. Like, the more people can play with Data understand cause and effect, the better they'll be as Data scientists and, you know, getting your hands on, you know, people who are trying to solve problems. And, you know, way back in the day, I don't know, this goes back 12, 14 years. Netflix offered one million dollar prize for the person who could come up with the best algorithm that predicted the next video that someone would want to watch based on their history. Does that sound at all familiar? Ah, it does. Yeah, yeah, yeah. I remember that. And they made that data available and they were really one of the first ones that I remember getting the Data myself. And, you know, my company's really early and we played with it and we built a model and and I was shocked at how predictive the winning model was because I looked later. [00:42:28] But, you know, it's you can always learn something when you're trying to really put your hands on a data set. And there's lots of problems in the world that have not been solved. Maybe it's not Netflix again. Maybe it's a homeless shelter that could use a lot more funding. That's a problem that's waiting to be solved. [00:42:48] So, you know, and it's only slightly limited by creativity when it comes to Data, when it comes to what you could do, you are really limited by your creativity. That's the only limit you have. And when you think of it that way, there is so much that you can you can do. [00:43:03] Totally agree. I totally agree. You know, none of us knows what the future totally holds, but I think it's really safe to say that Data science will be a great profession for many years to come because Data is always going to be an output of doing things. And learning from Data is always very important. And I don't I think we're many years away from a true A.I. that without any context can just step into a Data set and start solving problems and answering questions. You know, I think humans will have a role in that for at least for the foreseeable future. [00:43:39] Oh, absolutely. Absolutely. And guys, check the show notes. There's going to be a link to get a free sample data set from Analytics IQ. It's got five hundred rows, fifty columns. That's more than enough for you to do some real fun exploratory data analysis with, you know, do it and do it in Python. Tag me, tag Dave, tag analytics like on LinkedIn. I want to see what you guys are doing. Put it up on your tablet public's dashboard, tag me and tag analytics IQ. Let us know what you guys are doing with it because I'm, I'm super excited to see how you guys flex your creativity with this awesome data set and what you guys come up with. I've had a chance to play around with it and it's super fascinating. So definitely guys click on that link in the show notes and and yeah, tag me in what you do. I really, really want to see what you guys come up with. So thanks for that. Thanks for telling us about that, Dave. So you've been involved in some really amazing entrepreneurial initiatives. Do you have any advice or tips for anyone who's who's toying with the idea of entrepreneurship? [00:44:40] Definitely, as I said earlier, you know, make sure you have a competitive advantage of some sort. Without that, it is very difficult to step into any market and go, well, it's the same people just choose me instead of this other one for some reason. So that has to be there. And and admittedly, it's hard to be objective sometimes about it as an individual. So I always encourage people to have a focus group. And the focus group could just be family members who are painfully honest. For example, you run the idea by them in terms of traits to be an entrepreneur and succeeding. To be honest with you, the biggest one I've ever found is you just have that ability to blame yourself first when something goes wrong, if it's your personality to always blame others, even sometimes if an initial thought is maybe that's true, I feel like it's difficult to learn and evolve. And one thing that's sure for everybody who's ever started a business, and I think I would think this even applies to like an Elon Musk. Things don't always work out the way exactly where you think they will. Some things aren't going to go well, but you have to have the ability to objectively learn from mistakes and evolve and change. So to me, those are the biggest things. The common thing everybody says, but I definitely found it to be true is you'll need a lot more money than you think you will. So that's another thing that I found to be a fact. [00:46:07] Also, thank you very much for sharing that he had the accountability part. I think it's something, you know, I struggle with sometimes as well. And that's probably a big, big key factor for really being successful as an entrepreneur, because you are really like you are the one that you have to answer to at the end of the day. So thanks very much for sharing that. So, you know, in terms of data science, data analytics. Entrepreneurship in this covid world that we are in. What do you see as some some problems worth tackling that maybe an enterprising analytics professional can identify and maybe, you know, seize? [00:46:44] I think there's tons of room for innovation in all areas because Data science is such a broad subject. You know, people are working with all types of different Data, all types of markets. I've just found that there's still room for innovation. I believe the market still needs quicker and more effective ways to learn from Data feedback loops. So, for example, a company runs an ad. Well, you know, when can you get the data from that? We can learn from it so that you can make it better. I still feel that in most markets know that feedback loop takes still takes way too long. But in general, you know, I think whatever market you as a listener are in, I think there's lots of room for innovation. And I would just encourage everybody to step back and go, hey, looking past the specifics of this Data, how what would make things better? Like and, you know, sometimes it sometimes involves thinking about things that aren't realistic, but that can turn into things that are realistic. And so they're my general feeling is we're so early in data science. It's it's still a relatively early thing and that there are still a lot of opportunities. [00:48:00] Fluey. Thank you so much for that wonderful advice. I really, really appreciated it. So talk to us about the work that you're doing with Honey and. [00:48:07] Sure. So it's an orphanage has been around for quite a while. I, through a family friend, was fortunate enough to get involved with it. And we formed this partner and I formed a father when C three, which is the US legal entity, like a corporation for nonprofit in early twenty twenty. And you know, our mission is to to provide meaningful lives for these children. And it really helps you put things in perspective, especially those of us, you know, that are, you know, live and, you know, developed economies and you know where our problems are different than a problem. If you live in Uganda, in my opinion, especially if you're a young child without parents. So it does help you put things in perspective. I found it extremely rewarding. And be honest with you, my mission is to do my small part to to help make better lives for this small group of children, around one hundred kids, but to go really deep in terms of what we're doing for them. So that's that's my mission and what I hope to get out of it. [00:49:23] That's absolutely beautiful. And if people wanted to learn more or help you in any way with this initiative, is there a website they can go to? [00:49:30] Yes, it's Honi and Haven Dog. Our reach out to me directly again, you know, happy to educate people on the problem. There does. They don't, because believe me, the problem is massive. Our particular orphanage isn't the only way to get involved, but happy to educate people on what we're doing and how they could help. [00:49:51] Thank you very much. So that the last formal question before we jump into the random round. So it is one hundred years in the future. What do you want to be remembered for? [00:50:01] I'd say my first thing comes to my mind is to be a good father. Like I hope you know, my one hundred years of the future, my great, great, great grandchildren can go, wow. You know, obviously, he was a great dad. His kids turned out really well. You know, I think those of us who are parents, I think, you know, that has to come first. But other than that, you know, that that I was involved in starting and building something that in some form or another survived in it. You know, it helped on this journey along this Data science road to some extent. So, yeah, if I am remembered, I hope it's for being a good father, being a good boss, maybe making a small difference in some some lives of some children, hopefully many children over the years. [00:50:50] If you look into the future, it's absolutely beautiful. And I know you're well on your way to doing that, Dave. So let's jump in to our random round. First question here. What's your favorite stat that makes Data lovers think that's interesting? [00:51:05] You know, I mentioned astronomy earlier, the speed of light for some reason. I've just always found it totally fascinating. Three hundred thousand kilometers per second, the fact that anything could be that fast, yet the closest star there are still takes over four years at that speed for light to reach us. Like they're just kind of blows me away. The distance is the. [00:51:29] The scale in our universe, it really does put things in perspective great when you think about how far the distances is are and really how short your life span is compared to, you know, that distance. That makes sense. Absolutely. So as a fellow portioner myself, I've got to ask, what's your favorite model? [00:51:53] Well, the 9/11 for sure. You know, I wish I could say I owned, you know, not 9/11 three, which is a really specific Porsche. I don't. But, you know, the the 918 Spyder, I've only seen one of those in person. And I don't know if you have Harpreet that now. Amazing vehicle that costs a million dollars, which I'm against. But it is a super cool car. You know, we can all dream a little bit, I guess. [00:52:28] Yeah, yeah, yeah. I like the er the Panamera, I like big cars, I just like big and I mean it's not the fastest one but I like the Panamera and I've had two kind of mirrors. [00:52:39] They're great cars and they are big cars, no doubt about that. [00:52:42] The brand new Cheyennes are really nice. So my wife has got the mackan I'm more of like a Lexus guy myself. That's like my my go to car about the Porsche. My wife just has a push present because she just had her first kid just about nine months ago. That was one of those. Congratulations. Thank you. Yeah I'm nice but it's still like I feel because I mean I'm very much shaped like an avocado, so I feel like the person a little bit too small for me. But they're beautiful cars actually. I love them. They are. You have to appreciate the engineering, I think. So when do you think the first video to hit one million views on YouTube will happen? And what will that video be about? [00:53:24] Well, could you help me out by what's the record now like? I know. [00:53:28] So right now it is about eight point six billion views and it is baby shark. And before baby shark baby shark just overtook the Justin Bieber song. Esposito veterans, those are eight billion ish. [00:53:46] Wow. So it's still quite a ways to a trillion. You know, as the world gets more connected, you certainly have to be something that really resonates with a really wide variety of people. I've actually never seen either of those videos, although I do like that song. So I think it'd have to be something that really resonates in space. Something I can't predict would be like a cat doing something crazy or some, you know, song that someone in Albania puts out or something when it happens, if we're only up to eight billion, which seems just barely more than are humans on Earth, I've got to think it's going to be in the distant future. [00:54:24] You know, let's say fifty years from now, if it is the year thousand forty four right now and you're watching this, do your part and share this video. Do your part. Show the media. Let's help get help. Get this video. Two trillion views. So in your opinion, what do most people think within the first few seconds when they meet you for the first time? [00:54:47] You know, I think throughout my life it's just that I'm very tall. I'm close to six foot six and my son, who's eighteen, is six foot eight. So I know what I think every time I see him and I see him a lot is, wow, he's tall. So I feel like that's probably my, like, defining characteristic. [00:55:06] Six foot eight. That's tall. Yeah. Does he play basketball? No, he's a tennis player. Oh this. I know you're big into tennis as well. [00:55:14] I am. And my other son plays college tennis. So yeah, it's kind of a family thing. [00:55:19] So it's just become an impediment in tennis or is that an advantage? [00:55:24] Oh, it's helpful. If you look at the top tennis players, you have to go down the list quite a bit to find someone under six feet. You know, if they ever made the net shorter, I think being tall would not be an advantage. But with the height of the net, it's still a little bit of a tall person's sport. [00:55:40] That's interesting. I didn't know that. And that's another interesting Data analysis project to do. Go look at the the trends of tennis player Harpreet Sahota the last few years and you find. So do you think you have to achieve something in order to be worth something? [00:55:58] Oh, not at all. I think I think everybody is worth something. I think achievements are relative. You know, again, if if I was born in Uganda, perhaps owning a fruit cart where I could support my family is a much bigger achievement than if I was born in the US or Canada and had, you know, a middle class family and that I graduated from Harvard. So I think achievements are relative and I think it has to do with the distance someone goes in their. From where they started, but I think everyone has value, you know, like I said, value to me is tied to achievement. I think it's tied to innate goodness and, you know, and in caring and other traits like that, what are you currently reading? [00:56:48] And we talked about the astronomy book earlier, but is there anything else that you're currently reading? [00:56:52] Oh, that's a good question. Well, right now, that actually is what I'm reading. I tend to jump around quite a bit and what I read and like everyone else, I work on a computer a lot says there are periods of time where I don't really read anything. I may like to hear this. I listen to a lot of podcasts. Nice. And, you know, there's there's some in addition to your podcast, there's some great ones I've discovered that I listen to. And I feel like maybe that's the way I collect information. Now, for the most part, they switch podcasts. Are you listening to Do you know the sky? Lex Freeman? Yeah. Yeah. It came out of MIT. He's got a podcast. And the unique thing about his podcast is his episodes are usually around two and a half, three hours long. Yeah, but I find it to be really interesting. You know, he's got physicists on there. He's got philosophers on there, kind of a wide variety. So I like to said that's the one I've been listening to lately. [00:57:50] Yeah, that's kind of what I'm trying to do with my podcast. I've also had physicists, I've had philosophers and I've had authors of different books and stuff. So I'm definitely trying to imitate what Lex is doing and then trying to do to my own my own way. Hopefully one day I'll get Elan on my show like he did, but now we'll see. Yeah. What song do you have on repeat? [00:58:11] Well, you know, I guess I would put it this way. Like my favorite song ever is a song that most people won't know. It's called Given the Fly by Pearl Jam. And every couple of years I rediscover that song. And it's not one of the early songs. It's it's one that came out about 10 years ago, but it's about overcoming adversity. And I just think it's an amazing song. And like I said, I might go a long time and go, I'm tired of it. [00:58:35] But then I rediscovered again and got a bunch of times. So, yeah, something that has always meant something to me. I absolutely love Pearl Jam. I'm a huge Eddie Vedder fan, so that's awesome. [00:58:47] I've had the opportunity to meet him and I'll be older than he is, but very insightful guy. [00:58:54] Yeah, that's cool. So I'm going to jump into the random question generator. And the first question, when people come to you for help, what do they usually want to help with? What do they usually want to help with? [00:59:06] Well, I would say it's typically in my profession, typically people that work for me. So I think ultimately it almost always comes down to assessing Data to make decisions. So what I find is people come to me and say, hey, this is the facts around this, what should we do? And so, you know, ultimately, anyway, you look at it, it's it's looking at Data and making a decision. [00:59:38] I would thought it was a need help grabbing that thing on the top shelf demand that happens to but not around my household so much. He has run around that house. If you could have any superpower, what would it be and why? [00:59:53] Wow. You know, I think it'd be and this isn't a I don't think this is a Marvel superpower. I think it'd be really cool to know the future. Yeah. Like, I think that would be because I'm very curious and, you know, somebody who I want to live a long time for one thing, just to see what happens, you know. So I think that would be the thing if I could choose anything that would be like how to certain things and, you know, that that would be the one I would pick. [01:00:22] Yeah, probably the same thing for me. What's your favorite book? [01:00:25] My favorite book, you know, goes back to a book I read was a good Catcher in the Rye. You know, most people read it. I just it meant something to me when I was let's just say I was like thirteen or so when I read it. And while there are many other books that have meant a lot to me, that that probably is the one who I had to choose, one that I would go. [01:00:47] Yeah, that's that's the my top book. Yes. It's definitely a classic. In your group of friends, what role do you play? [01:00:56] The one who encourages people to travel and tries to get people to go on a bike trip, things like that. I'm also of Irish descent, so I certainly like to crack open a bottle of wine, so have dual roles. [01:01:13] So before the pandemic had hit us and we had all these travel restrictions in place, where was the last place that you had went to? [01:01:23] So I took a really interesting trip with my brother and I did a bike trip. From Prague to Budapest, wow, went through five countries, a really cool experience. [01:01:36] Yeah, that's awesome. Prague is absolutely beautiful, beautiful city. [01:01:41] I was of all the cities I went through and we went through four or five big cities, you know, Vienna, Bratislava, Prague, to me, stood out is like, wow, this is a place I definitely want to come back to. And we want I'm like, I can't describe it. It's just there's just something cool about it. [01:01:59] Yes. It's like black and gold everywhere. And was called the Charles Bridge. It's just so beautiful out there. [01:02:05] So like a really well organized city, I felt like you just hop on the elevated train. You can go anywhere. Great restaurants. Like I said, I, I can't I am planning to go back this summer if they allow it. But yeah, that's definitely a place I want to go back to. [01:02:22] So you took that bike. How long is that bike ride? [01:02:24] Well, it's not exactly sounds. It was a back road trip. So a give you a bike, you ride 30 miles or so. They take the bike from you and you know, you go to a hotel and you eat and drink, and then the next morning you're good. You're somewhere else along the way. So not like we hopped on our bikes with backpacks and rode that. Really. I understand it's I think it's 500 miles along a long, long way. [01:02:51] Yes. That definitely sounds like an interesting type of trip. I would definitely be up for doing something like that. Can't wait till things simmer down with this pandemic situation. Looking forward to something happen soon. Yeah. So, Dave, how can people connect with you and where can they find you online? [01:03:08] I think LinkedIn is a great place to find me. So, you know, Dave Kelly, analytics IQ, my email address, happy to provide it. Dave Kay, analytics, dash, IQ, dotcom, you know, definitely encourage anybody to reach out to me. Happy to help in any way I can. We are hiring data scientists, so, you know, thrilled if people are interested and would like to talk to us about that. [01:03:31] Absolutely. I'll be sure to link to all of the information right there in the show notes. I'll even link to the careers page analytics IQ. So if you guys are interested in being part of this company, definitely go check it out. Seems like an awesome place to work. And, you know, based on this conversation with Dave, you know, you could tell he's he's an awesome boss. So I definitely would would encourage you guys to look at that opportunity. Dave, thank you so much for taking time out of your schedule to come on to the show today. I really appreciate having you here. [01:03:59] Thanks, Harp. You know, I love your podcast. It's a thrill to be here, so. Yeah. Thank you. Thanks again for letting me.