Kirill Eremenko: Welcome to the SuperDataScience Podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur and each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast, ladies and gentlemen, and welcome to Australia. Today I’ve got a data scientist who I’m talking with who’s from my home country, Australia. Nic Ryan lives literally 4 hours away from where I am right now in Bundaberg. And on this podcast, we had quite a casual chat about data science and his journey, his career. So what you need to know about Nic is that he is a data scientist but he’s an external data scientist, like a consultant that goes into businesses and does data science work and at the same time, helps businesses with data strategy, with how they can use data on different projects, and things like that. So he does a mix of things all relating to data. And actually he also does mentoring of business employees to help them in data science. And what I love about Nic is if you go to his LinkedIn page you’ll see him surfing some waves in his background image and that’s what he’s all about. His lifestyle is very relaxed, very laid back. He’s built a life for himself where he can do data science remotely and occasionally go to visit his clients to help them out onsite.
Kirill Eremenko: So in this podcast you’ll learn how exactly he did that and perhaps some tips that you can apply to your own career to boost your success or maybe even build a similar lifestyle of remote work in data science. We talk about the whole notion of why data science is slowly becoming more and more popular in terms of remote work and also we talked about some of the projects that he did. He shared a couple of examples and we even talked about natural language processing, so you’ll hear some of those comments towards the end of the podcast.
Kirill Eremenko: All in all, a very chilled, relaxed podcast so get ready to explore the world of data science consulting and working remotely.
Kirill Eremenko: Welcome to the SuperDataScience Podcast ladies and gentleman. Super excited to have you back here on the show and today we’ve got a very special guest, Nic Ryan calling in all the way from Bundaberg, Queensland, Australia. Nic, welcome to the show, how are you going?
Nic Ryan: Very well, Kirill, and yourself?
Kirill Eremenko: Very well, too. It’s exciting ’cause very rarely it happens that not only we’re in the same time zone but we’re actually in the same country and the same state. We’re like four hours away from each other right now.
Nic Ryan: Yeah, and Australia’s a big country so you’re practically down the road. Four hours away.
Kirill Eremenko: That is crazy, yeah. That is so exciting. So you’re on the beach in Bundaberg. How cool is that? On a Friday morning.
Nic Ryan: Yeah, still some work to do but no it’s kind of been interesting thing. I grew up in Sydney, well the central coast of New South Wales, and bounced around between Sydney and Melbourne for work but yeah being able to get the lifestyle here in Bundaberg but also traveling to still Sydney and Melbourne and Brisbane for work is kind of good. It’s kind of like the best of both worlds.
Kirill Eremenko: That’s awesome and I think that’ll be a central topic for today’s podcast. The whole notion of doing data science remotely, of working in data science for living the dream, fulfilling the need to code and create stuff and do analysis, but being able to work out of the office. I love your comments about that it’s actually something that the world of development has already started embracing and there’s a lot of developers that work remotely from home, and slowly data science is moving towards that. So, I’m pretty pumped to talk about this on this podcast episode.
Nic Ryan: Yeah, and you do with your business as well, having that flexibility to … you mentioned as well you spent some time in Brisbane but you’re also in other places as well, and being able to move around and meet people and see new things and being able to have a real bird’s eye view of the data science community and what’s happening in different cities is kind of cool.
Kirill Eremenko: Yeah. So, you guys listening to this, Nic mentioned at the start before the podcast that he’s got some questions for me as well so I’m completely unprepared. We’ll see how this goes. It’ll be fun.
Nic Ryan: It’s a great opportunity to hit you up with questions ’cause as I said I’m a big fan of the work that you’re doing and the way that you’ve been able to democratize the data science education to help people out in their journey. I think it’s just awesome. And so yeah it’s a great opportunity to-
Kirill Eremenko: Thanks man.
Nic Ryan: Find out what you guys are doing cause it’s pretty exciting.
Kirill Eremenko: Thanks, man. Thanks. Alright, well, we’ll get to that but to get us started, to kick things off, tell us a bit about yourself. Where are you from and how did you get into data science? I read a little bit about your story and I think it’s pretty exciting.
Nic Ryan: Yeah, yeah. I grew up about an hour and a half away from Sydney, north of Sydney in between Sydney and Newcastle in Australia and I guess at school I was pretty good at math and I also played a lot of basketball. And so I always wanted to play basketball professionally. That was the dream, to go to the States and play basketball.
Kirill Eremenko: How tall are you?
Nic Ryan: I’m 6 foot 1, I was nowhere near tall enough, but that’s one issue. But I wasn’t good enough as well, Kirill, that’s the critical issue.
Kirill Eremenko: Oh.
Nic Ryan: You know like Muggsy Bogues he’s like 5 foot 3 or something and he’s able to do it but I didn’t have those skills. Yeah that’s a bit of a disappointment, bit of a letdown, but my career advisor at school said “hey there’s this thing called actuarial studies and you can do that at university” so um-
Kirill Eremenko: Very similar to basketball. Second thing behind basketball. Next best option.
Nic Ryan: Well yeah, next best option and it’s kind of funny as well here because I still play basketball a couple times a week, my dad still plays basketball actually.
Kirill Eremenko: Your daughters play basketball.
Nic Ryan: My daughters, my dad, the whole family, everyone plays basketball. My dad’s 63 he’s still playing.
Kirill Eremenko: Wow.
Nic Ryan: Basketball. And I still play a couple times a week, but yeah more for fun, but a lot of the guys here actually, are playing at a high level. A couple of them have gone over to college in the States which is pretty amazing from a small place like this. But anyway, I digress.
Nic Ryan: I did actuarial studies and I ended up working for insurance companies back in Sydney and Melbourne and that was really in the early days where we started to fiddle around with some aggression modelling and we started to do some special smoothing and some mapping and early fraud detection, and even some of those sorts of models. And it was really the statistics and what was data science machine learning that I really got into but I didn’t really know how to code. I didn’t really know programming, and that was my weakness, and so I just worked on those skills, and over time this institute just grew around me. So I ended up working in banking building credit risk models for banks, and doing some consulting to local lenders in Australia, building their risk models that were mainly regression-based models, over many years. Then as well, working with some different start ups and the natural thing was to then start my own consulting business.
Nic Ryan: Where I fit in really is, all of the companies they’ll have their senior leaders and they’ll have their staff but I fit in as a technical manager there, compared to someone like a Deloitte or KPMG the rates I charge are very reasonable and it’s something that … the companies that have say 20, 50 people, and they’re interested in AI, they may not be able to afford some of those other guys but where I come in-
Kirill Eremenko: Gotcha. So you found your own niche.
Nic Ryan: Yeah, yeah, it rolls off the tongue, but that took a couple years of trying.
Kirill Eremenko: That’s actually what I wanted to dig in. First of all, you totally skipped the exciting part about the two hour train ride and how you fell into data science. Tell us a bit about that and then I have another question for you.
Nic Ryan: Yeah, that’s right. So that actually does take me back. Sometimes you brush over the more painful parts of your life, kind of keep them in the back of your memory there. But, yeah, I was pretty young, I had the kids, and so I was on the last part of my actuarial exams and I knew that actuarial wasn’t for me. I just knew it wasn’t something that I was really passionate about, and I think-
Kirill Eremenko: What is actuarial for the sake of those that might not be familiar with this profession?
Nic Ryan: Yeah, it’s more about predicted modelling and maps and stats for insurance companies, really. So to predict [inaudible 00:10:04] Some that they’re reserving to be able to predict when clients are going to occur and how much they need to put aside and some of it as well is around pricing of insurance products, as well. But they’re kind of like the people that do the math and stats for insurance companies.
Kirill Eremenko: Yeah and what I like about actuarial is the whole … they deal with demographics and population statistics and how long are people in this age group, what happens when they transition to retirement? There’s a bit of this whole human or social component to it. That’s what makes it a bit of a different profession than just pure plain statistics that you’re just dealing with numbers and mathematical equations.
Nic Ryan: That’s right, yeah. And there is a certain matter to main expertise that you have to know in any field like that and especially in insurance, yeah, you need to know, be down with some of those. And things about the different policy rules, the different insurance products as well, so yeah it was interesting. A lot of it as well, even back in those days, I think it was a lot of spread sheet work and so it almost felt at times like glorified accounting. Not that there’s anything wrong with that but it was … it certainly has changed now where there are data science professionals attached to actuarial teams and they’re doing some really cool stuff. But for me at that time it wasn’t something I was super passionate about.
Kirill Eremenko: Yeah.
Nic Ryan: So, to each to their own. So then I was moved over to banking and I did get really passionate about building risk models for banking which is weird, but I really thought it was pretty cool. And so I had a pretty long commute from on train, we’re talking close to a couple hours each way, door to door.
Kirill Eremenko: Wow.
Nic Ryan: Each day. So yeah, 4 hours. Four hours where you’re not around seeing your family and seeing the kids and everything, so I was stuck on the train and I thought well I can either get sleep or I can use this time productively. So there’s probably about 2 and a half hours a day which I had to study and it was the early time of Coursera, and they had some of the moocs out there and I started learning the programming language and did some of the courses through John Hopkins University, like Jeff Lake, and those guys. And also, the machine learning course [inaudible 00:12:30] and just a few others, and I thought “wow this is special. I’m really enjoying this.” I really had a passion for it.
Nic Ryan: And so, within a few years, I got a phone call to move to the Gold Coast and to essentially head up data science for a start up there. And so the time from learning to something significant happening was pretty short-
Kirill Eremenko: Interesting. Out of curiosity, you weren’t actually … people don’t find you through learning unless you put yourself out there. Did this happen through your Linkedin page? Did this happen through something that you shared? Did this happen through a connection? How did that phone call happen?
Nic Ryan: Yeah, I wrote a pretty early blog, it’s just a recruiter happened to stumble upon it and I really think that putting yourself out there, it is scary but it’s definitely something you need to do. And it’s absolutely essential for my business to be out there, ’cause that’s how people find me. I mean you can knock on doors and you can try to say what you can do but showing what you can do and having people come to you is usually a better equation.
Nic Ryan: And so that’s what happened. They were looking for a head of data science and they looked for 4 years and they couldn’t find someone. They had pretty simple criteria. They wanted someone who could do the technical work and someone who could also talk a bit, and it’s really hard for them to find both. And so, it’s extremely hard to hire people that can’t communicate as well as doing the work. And they even tried as far away as the U.S. and other places as well and so when the recruiter said “there’s a guy in Sydney who you might want to have a chat to” and then I had a plane ticket and off I was. And you’re on the Gold Coast now, it’s a great place to live-
Kirill Eremenko: Yeah, fantastic place. It’s a dream place. It’s called the Gold Coast for a reason.
Nic Ryan: That’s right. Where you are now, Southport is where I was living as well, so I would go for a surf before work and then go into the office. I’d skateboard into the office. I’d commute by skateboard, so yeah not very professional showing up with a skateboard under your arm but there you go.
Kirill Eremenko: Yeah. That’s pretty cool. That’s pretty cool. Actually, yesterday I was walking on a bridge in Gold Coast and this guy in a not too formally dressed, but in a shirt, looking smart, on a skateboard, probably your protégé.
Nic Ryan: That’s right.
Kirill Eremenko: Taking off to work.
Nic Ryan: My little brother’s actually a lawyer and he actually works with my dad who’s also a lawyer and my dad was driving to work and my brother’s really keen on electric skateboards, so he’s weaving in and out of traffic on an electric skateboard and my father thought “who’s this idiot weaving in and out of traffic. Oh, it’s my son.”
Kirill Eremenko: Oh, that’s too funny. Okay, gotcha. So, that’s a great comment. Even before you started your business … I completely agree, when you have your own business it’s important to put yourself out there so people can find you and connect with you and know that you can help them out. But even before you started your business, putting yourself out there, writing that blog post was an important way to put your foot in the door and actually for that recruiter, it showed them that you’re a person who can potentially communicate and maybe take on this role. So just for those listening out there, it’s a great note to take. It never hurts. It is scary to put yourself out there and share some of your thoughts or learnings but it never hurts. What harm can it do? It can only lead to good things, right? Otherwise you would’ve never gotten this phone call, never moved to the Gold Coast.
Nic Ryan: Yeah, that’s right. And for you. How did you get started? In a similar way, also putting yourself out. ‘Cause for you, you were working for Deloitte and you were doing some consulting work and then I’m just curious about how you got into doing the courses-
Kirill Eremenko: Good question. Here come the questions.
Nic Ryan: Sorry.
Kirill Eremenko: No we’re good. For me, I worked at Deloitte for 2 years and, amazing company. I couldn’t recommend higher, the professionalism and excitement you experience when you’re at Deloitte, different types of projects. But if there comes a time when you … life changes or you feel that you want to move to the next level or something else is exciting … for me, that was after 2 years I felt that “alright I’ve had enough of all of this.” I worked on probably a dozen or more different industry projects, learned so much, and grew very fast, and then I kind of hit a ceiling in terms of my growth, and I decided “alright I want to do something of my own. “
Kirill Eremenko: And from there I started searching for different options. How can I start a business or how can I become … make money in a passive way, or how can I help people do certain things that I’m passionate about? And stuff like that. And one thing led to another. I started putting out courses on things I knew. And at the time, I knew really well a lot about forex market and how it works and financial instruments and stuff like that, so I started putting out courses on that topic. And again, similar to you, because I started putting myself out there, found this platform, Udemy, which I found completely randomly through an eBook that I was reading about another instructor who’s course I was taking. Anyway I started putting out these courses. And that went well. It turned out that I’m pretty good explaining complex things in simple ways and Udemy themselves reached out to me and said “hey what else do you know?”
Nic Ryan: Yeah.
Kirill Eremenko: And I said “data science” and they said “oh how about you create some course on data science,” and that’s how it all started. But similar to your story, had I not started, had I not gone through that painful, fearful exercise of releasing my first course on writing algorithms for trading stocks, not stocks, currencies. If you listen to that course now you can hear how shy and how timid I am. I’m nervous, almost panicking on the microphone. But had I not done that, had I not pushed myself through that experience, then that would’ve never led to that, as you say, phone call. For me, it was an email from Udemy saying “hey man, let’s do some more stuff.” And from there, turns out that data science is something that people need. Lots of people have, since then, have been able to learn from the courses that I’ve created. I’m really excited mostly about that part. That, that’s where it all led.
Nic Ryan: Yeah, I think as well with what you guys are doing now, you’ve got the SuperDataScience Platform the 2.0, that you’re doing, and so that’s access to all your courses, and for what is I think a pretty normal figure, about 150 bucks a year as well so it almost feels like you’re graduating again from Udemy in a way.
Kirill Eremenko: Yeah, yeah.
Nic Ryan: You’ve got something pretty exciting there, too.
Kirill Eremenko: You always got to grow and develop right? I find in life, every new experience is like a step, and then you learn and you’re like okay, cool, that’s great but as soon as you feel there’s something more, some new way you can go, that’s really cool, and thanks for mentioning the SuperDataScience 2.0 which we just launched.
Nic Ryan: Yeah.
Kirill Eremenko: Very very excited about that. By the way I wanted to ask you. How did you find out about SuperDataScience and things that we do?
Nic Ryan: No, I mean, yeah through Linkedin. Linkedin’s actually been a good way to … for me, I’m quite isolated here. There’s no other data scientists around. It’s really just me in the area, and so I probably have to drive down the road to where you are to find someone else to talk to. With social media, with Linkedin, with even Twitter, with those sorts of things, you can really be connected to the industry wherever you are. And so, just keeping tabs on what’s happening. I’ve done a couple of your courses as well, and I really love them as well.
Kirill Eremenko: Thank you.
Nic Ryan: And in particular, the AI course as well-
Kirill Eremenko: Which one? I think we have five of them.
Nic Ryan: Put me on the spot, I don’t remember, but it’s the one with Hadelin and he does … I can’t remember what it’s called. AI or something like that.
Kirill Eremenko: Is is the recent one or is it AI, artificial intelligence A to Z, or is AI for business?
Nic Ryan: It’s an A to Z one.
Kirill Eremenko: It is. Oh, okay. The fundamental one.
Nic Ryan: Yeah, the fundamentals. But what I liked about even those courses as well is that some of the … if you look at YouTube tutorials, you’re kind of passively coding and you just kind of look up and code some more, look up and code some more, but with what you and Hadelin did is you’re like oh okay, here are there papers. Go off and read some of these papers.
Kirill Eremenko: Yeah.
Nic Ryan: Or, here’s something else. And it’s kind of like it’s putting the ball back in the court of the student to take responsibility for their own learning as well, and you’re kind of acting more as a guide rather than just telling them “hey do this, hey do that.”
Kirill Eremenko: Yeah.
Nic Ryan: So, that’s pretty cool. And some of the other courses going that way as well-
Kirill Eremenko: That’s awesome. And what we … I don’t want to make this podcast a promotion of SuperDataScience 2.0 but I think its important to mention that what we aim to do with this new platform, that’s why we’ve been developing it for two years actually.
Nic Ryan: Wow.
Kirill Eremenko: Yeah, it’s been a while. But the main thing here is and it’s constantly in development, constantly improving, but one of the things that we’re releasing very soon is the gamification component, where not only do you take the course and as you said you have exercises or you have papers to read and things like that, but actually as you progress through a course, you get certain badges, and unlock achievements, and get points, and that’s … even though it might sound very childish or not for adults and things like that, but actually it is very cool to see your own progress and feel like oh that’s awesome. I unlocked this level in my education. And the goal of that is not to turn it into a game but to actually help people get into it, motivate them.
Kirill Eremenko: How many times have … I’ve done this plenty of times. I sign up for a course and I don’t actually, I take one tutorial, or I don’t even look at it or anything like that. So, but if somebody helped me get into it, somebody helped me get started and get momentum and I realized how powerful this is for my career and for my personal growth, then I would continue going. And so that’s one of our main missions now is how to actually help people select the right career paths and courses for those career paths, but then actually motivate and inspire people to keep with education. Stick to it and hold themselves accountable to it.
Nic Ryan: Yeah, and I think as well, for anyone, well anyone that’s listening and if they are taking some responsibility for their own education, if they are trying to learn new things, immediately that puts them in the top small percentage of people out there.
Kirill Eremenko: Yeah.
Nic Ryan: Most people, and I’ve known this from managing teams and that sort of thing, but very small percentage of people, whether they’re software developers or data scientists will learn outside the job. They tend to … a lot of people, their learning is their 9 to 5 job and they don’t want to do anything outside. So, if someone is actually taking that initiative to sign up for courses and to do it, they’re already ahead of the game. And that’s the thing as well, when I used to hire people, it’s just is this person as passionate as I am? I’m working with a guy over in London, really nice guy actually. Martin Paver. His company’s called Projecting Success and it’s a start up and he’s applying AI and machine learning to project management data analytics, so really quite exciting. Really quite changing the project management space. But, I was speaking to him a while ago and he said words to the same effect, that when he’s looking at hiring people, whether interns or more senior people, he’s looking for evidence of that passion. And so he’s looking for … you could get someone and they could’ve spent 100,000 US dollars on their education, which is fine, but it’s really are they going to hackathons, are they looking at moocs and doing online courses? What’s the evidence that I can see that this person is passionate about the field?
Nic Ryan: So, for me as well, even if they don’t have a high school education, that’s kind of irrelevant if they’re super passionate about the field, because anyone with that passion and that drive and if you guys are helping them to get that passion and drive and keep them on track like anyone can do this stuff. And they just gotta be pumped about it.
Kirill Eremenko: Yeah, totally agree. That’s very valuable. For those who were at DataScienceGO last year, Ben Taylor actually talked about that, that companies want to hire people who are somewhere between passion and obsession about a topic. You gotta be passionate or even maybe obsessed to certain extent, and then it’s a no-brainer. Companies will want you on their teams.
Kirill Eremenko: Alright so that’s a quick digression there.
Nic Ryan: That’s right.
Kirill Eremenko: Side route. I wanted to rewind back and this is the second question I wanted to ask you. So you told us about the train ride and how you put that to use, which I think is very inspiring. Tell us the other thing that you said was “it was a natural step to start my own business.” It might seem natural for you looking back, but trust me, it’s not a natural step to start your own business. The way I remember it, it was a lot of trial and error, a lot of fear, a lot of “how do I do this? Will I succeed? Will I not?” So, how did that all happen? You moved to the Gold Coast, you were working there. Walk us through how you went from that point in your life to actually becoming a consultant and working for yourself, working remotely.
Nic Ryan: Well again, I guess you don’t know the history, so my wife’s mother moved to this area, Bundaberg, like 18 years ago, so they’ve been here for a long time. I always thought this was a nice place to retire and eventually die. And so, I thought my ashes would be scattered on that beach or something it’d be nice. So I was-
Kirill Eremenko: Long term planning.
Nic Ryan: Long term planning, yeah that’s right. It was always gonna be here. So we had been holding here for a bit and we really liked the area and it’s nice and in terms of weather it’s the same as Hawaii, like all year round, it’s just awesome. And so, people here are very friendly, and really nice.
Nic Ryan: I was working for that company on the Gold Coast and I was managing a team that was based in Kiev and also got to go to Kiev, which is a lovely place in Ukraine. It was fantastic. And also a team on the Gold Coast as well and just pushing a bit hard because there was a bit of pressure in that job I will say. And, responsible for a lot of people. So at the expense, I didn’t get the balance right, you know I didn’t get the balance between family and work and life right. So it was not great. So I didn’t see my kids all that much and I was almost going back to when I was catching that train and commuting long hours to Sydney. So I thought well, something has to change here, because my kids, and it’s actually probably an important point for people as well, your colleagues and stuff are good and you should be nice to your colleagues, but really your family and your kids are critical, because in 5 years time they may or may not … you know your colleagues may or may not remember you but in 80 years time, your kids will.
Nic Ryan: And so it’s really important to get that balance between family and work right. And so my wife, we went camping in the middle of nowhere, about 3 hours west, in the middle of Australia and then she just said right “we’re moving here.” Cause I was relaxed, we’re looking at the campfire and she said “we’re going to live in Bundaberg,” and I said “well there’s no work there,” and she said “well tough you’ll find some.” So that was it. And so she picked a house that she liked, she bought it and that was it.
Nic Ryan: So, I kind of had to, and that was pretty good motivation as well. And I always managed change. I always liked mentoring junior data scientists and it was a natural thing to be able to manage up as well as manage down to some of the junior data scientists and it is something that I really enjoy. In the same way I enjoy taking some of the kids around here for basketball and coaching them and within 2 years, some of them are playing representative basketball. I think “wow, that’s kind of cool to see.” So, I’ve always liked that, always liked taking someone who’s fairly green, a novice and being able to help them out on their path, whether it’s basketball or even data science. So it seemed like pretty natural thing to do.
Kirill Eremenko: Gotcha. So, very interesting and very … overnight transition almost, and would’ve been challenging but at the same time exciting to make it all work. Tell us a bit about remote work. So, you moved to Bundaberg. How do you find your clients? Or how did you find your clients at the start? How did you set up this whole system for yourself where you on one hand are relaxed and at home, but at the same time you do have a stream of work coming in?
Nic Ryan: Yeah, I mean I was getting paid pretty well when I was working in the city so I did have a bit of money which I could go on for a while. So it wasn’t that immediate pressure to find the next dollar. Also, not having a mortgage or anything like that. You can live pretty cheap here and we’ve got a fairly big lock of land, you know 7 chickens and grow our own fruit and vegetables, so my expenses really are quite minimal.
Kirill Eremenko: Okay, gotcha.
Nic Ryan: So I didn’t really need much money to be honest. It was only just my expenses are pretty low. So I could’ve kept going for a bit while. But what I was doing actually before then was just writing, ’cause I was … well, to start off with there was a little start up company in Melbourne that I was working for and so that was good. I just started writing about some of the things which I wish I’d learned when I was learning data science or some of the things I was seeing or just some of the thoughts, because if you can picture it, I’m in a room by myself writing code, so it’s a little bit isolating, but Linkedin as well has really helped to build that community.
Nic Ryan: Eventually people started approaching me for different ad hoc tasks and some for even longer engagements, so I typically might do a few hours here, a few hours there for different people, and there’s actually a really great AI consulting company that’s based in Brisbane called Blackbook AI and Thuy Lam, he’s a friend. He’s a really nice guy but he reached out to me. He’ll just book a plane ticket for me and say hey Nic, we need you this day, and I’ll be flying into Brisbane. Working with them is wonderful and they’ve got a great mix of junior and senior people and so being able to mentor and help out those guys is just awesome.
Kirill Eremenko: Oh okay, gotcha.
Nic Ryan: And so that’s the … and also help them on projects when they need it and a combination of meeting clients and also doing some of the data science work is kind of cool. And so it really is a mixed bag of stuff that I do, and it keeps it really interesting.
Kirill Eremenko: Very cool. Okay, so you not only do the data science work but you actually help them build teams, help them get their staff on track and consult the executives and things like that. So I think that’s a very exciting space to be in, especially now with the boom in data science and artificial intelligence.
Nic Ryan: Yeah, and I’m pretty picky about the people I work with. I think you have to be and you have to be in it for that long term relationship and so those two places that I mentioned, Projecting Success over in London and Blackbook AI in Brisbane, just great people. Just great people to work with. And so, you do have to be selective and kind of in it for the long term. It’s also a bit more of a pipeline and there’s some really cool people that I’ve met in Melbourne last week and also on the Gold Coast as well, this other little start up company. Some of these things still in the pipeline, but yeah, they seem like really good people as well so it’s cool. It’s excellent working with great people.
Kirill Eremenko: Awesome, awesome. So, tell us a bit about projects. You mentioned that you’ve been doing quite a lot of different … working on a lot of different projects, different companies. Are there any case studies that you can share with us of recent work that you’ve done without disclosing clients, or disclosing any sensitive information, just for the sake of maybe what kind of tools you did or used or what kind of changes you helped the client implement in terms of their staffing, in terms of their strategy, and whatever else that you were working on recently?
Nic Ryan: Yeah I think the one that’s really quite … there’s a few and obviously I’m going to be careful about talking about it too much, but what I think is really quite incredible that Projecting Success over in London are trying to do, is essentially, they’re trying to build a credit bureau, something like a credit bureau but for project management. ‘Cause what they’ve seen is a whole heap of different companies and different industries are doing different projects and they’re not necessarily learning lessons from those projects so they’re repeating the same mistakes. So what you’ve got is a whole bunch of spreadsheets and emails and word documents to manage projects. And so they’re about digitizing that information and storing it in databases and eventually all companies pooling their data collectively to be able to … so say I’m an agriculture company and I’m doing an IT Infrastructure change. Well maybe I can learn some lessons from another IT Infrastructure change by some other company maybe doing manufacturing or doing something different, but it’s still an IT project.
Nic Ryan: And so what they’re looking at doing is building really cool databases to collect that information. And to work out multiple parts and that sort of thing for projects to keep them on track and to minimize spend and to minimize the chance of it going around over time with no budget, which is a really inspiring kind of thing. So there’s one that completely changed that project management space, which is cool thing to be a part of.
Kirill Eremenko: Very cool. And so what’s your role in all of that?
Nic Ryan: So, for me, I’m helping out with … well they have a combination of interns and they have also a CTO who I help out as well, and even Martin who’s the founder, just helping him with strategy and advise as well, so it’s everything from meetings with web developer and seeing the different states of progress of the applications through to … yeah just different strategy and advise-
Kirill Eremenko: Interesting. So you’re not, in that specific situation, you’re not doing any data science work per se. You’re not bringing to life any models, running any logistics or other types of regressions, doing data cleaning. You’re actually acting as a data strategy consultant or helping them understand how data can be used and applied in the tools that they’re creating. Is that about right?
Nic Ryan: Yeah, for that project. I mean they do other things as well where they’ve got a really good … companies may try for a … [inaudible 00:37:39] They built a tool which is again a machine learning model which predicts the chance that someone’s gonna win and how much and what position they’re gonna get in that particular bid. So, and that’s more hands on stuff that I’ve done.
Kirill Eremenko: Yeah.
Nic Ryan: So, again, it’s a mixed bag of different work.
Kirill Eremenko: Okay, gotcha.
Nic Ryan: Yeah, and even with Blackbook, even. Yesterday actually I had all the tools and again, bit of a pipeline script that they’re just going to plug into production for a tool that they’re creating, so again, I keep it-
Kirill Eremenko: It varies.
Nic Ryan: Yeah, it’s good seeing different things. It means that you’re always learning as well, and so the imposter syndrome is very real. People will often ask, some of the junior data scientists will say “oh I don’t know how to do this.” I’m like “oh just Google it,” that’s what I say. It’s hard. You’ve probably found that as well when you started as well. You were thinking “oh, do I know enough? Am I good enough?” And all that sort of thing.
Kirill Eremenko: Yeah. I heard recently that this question: “Am I good enough?” Or that “I’m not good enough” is the most frequently asked question in the world that people ask themselves. And it’s the cause of the most misery in the world as well that people think they’re not good enough and therefore they’re going to spiral into depression or they don’t go for opportunities. They think … that actually could change their lives or they don’t go up to the love of their life … to meet the person that might become the love of their life, and things like that, so yeah, definitely I’ve asked that question a million times. For sure.
Nic Ryan: Just one of my questions. I’m gonna shoot the questions at you now. What do you do for fun? Like I do a combination of basketball and surfing and skateboarding but what do yo do for fun? Do you have something that you-
Kirill Eremenko: Oh man, good question. I like sports. I notice this about myself that … I was talking about with one of my colleagues at work recently and he likes what are they called? Going to a museum and looking at paintings and things, so something more for the soul. I like those things as well. I like listening to a classical music concert or going to a museum, but if I had the choice I always go for experiences. I go for, like recently, last week I went rafting with my dad who’s also in his 60s. He’s 66. We went rafting in New Zealand and we dropped down a 6 meter waterfall. That was really cool. I like scuba diving. Monday this week I went for scuba diving in Byron Bay.
Nic Ryan: Oh.
Kirill Eremenko: Yeah, like one of my top 5 scuba dives ever, so you know … I saw a bird under water.
Nic Ryan: What?
Kirill Eremenko: I saw a bird under water. It was like the last thing I expect to see, one of those birds, you know how when you see a bird from the surface then it ducks to go do something?
Nic Ryan: Oh, yeah yeah yeah.
Kirill Eremenko: And you never know what they’re doing. I actually saw it doing its thing. It was swimming around, it was like 40 seconds underwater, went to this rock, looked under that rock, then went and chased this fish, went back up, got some air, came back down. Ridiculous, man.
Nic Ryan: That’s incredible. Yeah, that’s something I mean to do is go scuba diving cause even on my street there’s world class dive spot that you can go scuba diving. I haven’t been. I’ve been here two years. I still haven’t gone there.
Kirill Eremenko: Man you should do totally do it. It’s like a whole different world under the surface of the ocean.
Nic Ryan: Yeah, I mean you’ve probably done the scuba diving course or-
Kirill Eremenko: Yeah, yeah, yeah. You gotta, if you want to establish yourself you gotta do the PADI or some other certification like the open water advanced, and other levels that you might want. But it’s fun. I like experiences basically. I like sports as well but something new, something where that’s physically challenging and that you feel engaged in and then after it you’re like “wow that was so cool.” Those types of things.
Nic Ryan: What sports do you play?
Kirill Eremenko: Oh, good question. This is a get to know Kirill podcast.
Nic Ryan: No no, sorry. I’m just-
Kirill Eremenko: Sports, well, what do I play right now? There was something that I did … oh I was doing Brazilian Jiu Jitsu recently but before that I really like Taekwondo. I did Taekwondo for 9 years, so martial arts mostly, and I don’t know it’s kind of like the challenge, that adrenaline that you get and yeah. And the whole concept of getting better and learning new techniques, so I would say martial arts.
Nic Ryan: Yeah, awesome. Yeah it’s something that both my daughters do as well is Taekwondo and my wife actually represented the state in Shotokan.
Kirill Eremenko: Wow
Nic Ryan: Yeah, I can’t do anything.
Kirill Eremenko: Wow, man.
Nic Ryan: Everyone in my house can beat me up.
Kirill Eremenko: Oh wow that’s crazy. That’s so cool. I heard Shotokan, that’s karate right?
Nic Ryan: Yeah, it’s a karate.
Kirill Eremenko: Pretty brutal type of karate.
Nic Ryan: Yeah, yeah, she can defend herself pretty well.
Kirill Eremenko: Awesome, man. Alright well let’s jump back in. What were we-
Nic Ryan: Sorry, got you off track.
Kirill Eremenko: All good, all good. We were talking about how … oh this is what I was gonna say. You wrote, and I really like this comment that you made in the notes for this podcast, that you think data science leaders should remain on the tools maybe floating in between projects as needed, and then you continued “if you take your best data scientist and make them a manager, you probably end up upsetting them if they don’t have real work to do.” And it sounds like that’s what you’re doing. You’re floating between helping companies with their strategies and tactics around data, what they can do, and mentoring people, but at the same time you’re not losing your grip on the actual applied data science where you get to code things in python or r and create things in regression models and what not.
Nic Ryan: Yeah.
Kirill Eremenko: What are your suggestions in general to people? How do you maintain that balance of the two?
Nic Ryan: I think, looking back in retrospect, when I was in charge of a fairly big team, and a lot of what I did was meetings, and admin and approving sick leave, and all that sort of stuff, and that’s … I don’t know, I kind of like you, you really enjoy the work. If you are passionate about it, you do love the work and anything that’s gonna take you further away from that work completely isn’t great and so you … I think as well, sometimes a lot of people, they seem to want to become a manager just because it means more money and I think that’s the wrong way to reward people ’cause a highly skilled technical person I feel should get paid just as much as a manager because this is something they’ve invested in. They’ve invested in themselves.
Nic Ryan: For me as well to be able to relate to what a starting data scientist is doing, it’s really important to keep on the tools and even in my spare time to be looking at what’s coming up and what’s out there. Actually that’s something true for you as well. How do you keep up with the industry? Cause it seems to be hard work just keeping up with what’s coming up.
Kirill Eremenko: Yeah.
Nic Ryan: But I think as well if you’re on the tools you’ve got a better chance of doing that.
Kirill Eremenko: Yeah man, totally agree. But for me, I have the luxury of … the content that I create, the courses, I have to still relate it to the technical stuff. If anything I’m not getting enough practical applications of data science, like solving industry problems with existing data science tool. But if anything when Hadelin and I are creating courses, we’re pushing the boundaries of data science.
Nic Ryan: Oh, yeah.
Kirill Eremenko: We’re looking at the most advanced tools. But I agree with you. I would love to be able to do more commonly accepted python or r models that currently dominate the world rather than just only focusing on the ones that are on the fringe that will become the dominating ones in a year or two. But you know, as you said, there’s always … you can’t have everything, right? You gotta balance it out.
Nic Ryan: No, I meant with … obviously your courses you’re quite technical but I mean there’s just so much technical content that’s coming out. It’s really hard to keep up. But I mean, you guys are massively on the tools and the way you decompose quite important and difficult concepts down for people shows that you are really on the tools and you’re really into the nitty gritty of the content which is incredible. But just the amount of stuff that’s coming out, I think it is good for me to be across different teams and different cities doing different things to be able to see some of the commonality between what people are working on, what they’re concerned with. And even just for some reason right now there’s a whole heap of natural language processing tasks that are happening all over the place and that wasn’t the case even a couple of years ago, so it is kind of good to see that. But then there’s so many libraries and the tooling for data science is just getting better and better all the time, so it’s often a question of working out what’s out there.
Nic Ryan: Even just yesterday I was working on trying to find a way to detect emotion from text, whether someone is angry, upset, disappointed, whatever, and I was thinking “how do you do that? Is there some machine learning API from google or” … then all of sudden there’s this little odd library that we’re doing. I just think far out, it’s hard to know. You pretty much have to read the internet sometimes to find out all these different things that are out there. There’s a lot out there and so it’s almost a job just keeping up, I think.
Kirill Eremenko: Yeah, no I totally agree. I actually read your post on Linkedin about natural language processing and it’s interesting how you mention that a lot of companies still talk about structured data but around 80 to 90 percent of the world’s data is in unstructured format, and I agree with you, it’s so much media. There’s so much video, there’s so much audio, there’s so much people writing stuff, like texts, blog posts. How often do we actually deal with structured data as humans? Not that often. Not compared to how much unstructured … like we’re talking right now, people listening to this podcast, that’s all unstructured data, so it’s important for data scientists to know how to process things like natural language.
Nic Ryan: Yeah, even in applications like credit based lending. I used to find it extraordinary that a lot of the banks and lenders were sitting on transactional data from bank statements and they didn’t incorporate those into credit risk models and so that’s a gold mine to unset and that’s something that I did and I was working for the on star company on the Gold Coast, was building natural language processing for bank statement data which was incredible. Incredibly rich data source for lending.
Kirill Eremenko: Interesting. Very interesting. Tell us a bit more about that. You obviously have a lot of experience with natural language processing. What are some of your tips or maybe some of the tools that you use and maybe some of the mistakes that you’ve made that you can help other people avoid?
Nic Ryan: Yeah, plenty of mistakes. I think often people will try to, not just with natural language processing, but with any kind of task, if you saw before we started, you can tell by my haircut and my t-shirt that I’m a pretty simple guy. So I usually like to start very simply. And in a natural language processing task it is about cleaning the corpus. It is about making everything, just the basics, everything lower case, stop words, looking at frequently occurring words, looking at infrequently occurring words, you know stripping out any numerics, and just cleaning the corpus and then really eyeballing the data. Even just some of the plots, some of the word cut outs, and some of the things could be inside of what you’re going to be looking at as well.
Nic Ryan: And so initially when I start, I’ll go for a pretty simple bi-gram or tri-gram, which is just 2 and 3 word features of words just to create context. So if you’re talking bi-grams, you might have say … you know that coffee shop is usually the example I give called Gloria Jean’s. If you’re looking at uni-grams or single words then you’d say, any time you’d hear jeans you’d go off to fashion-
Kirill Eremenko: No.
Nic Ryan: Whereas Gloria Jean’s is coffee shop, Gloria Jean’s. So having those two words as a feature in the model is gonna be important to give you that context and even three words sometimes for different company names or whatever.
Nic Ryan: And so, I would also just try with a simple regression model. So something like a multi class logistic regression model would be the very first thing that I’d try would be the very simplest thing. And to see how that goes and then if anything, that forms a baseline that you can then use to improve with better models. And in anything I do, I’ll always start off with a fairly simple baseline and then in track and develop on that. But getting the end to end pipeline and getting something going is more important than getting something optimal. And you were the same as well. If you’re working on a machine lending task, you might be able to get 80 percent of accuracy within a week or two, but to actually get 85 percent may take you months, or it wouldn’t be possible. So, yeah, I’m a big fan on building something simple and then intrading on that.
Kirill Eremenko: Yeah, that’s very cool. Get going and then improve, improve, improve, rather than just have the end ideal scenario in mind and try to strive for that forever.
Nic Ryan: Yeah, that’s right. And you never really learn until the system is in production and then you probably want to slowly get it into production as well, so test it out another couple percent of your population, make sure it’s doing the right thing, stop monitoring for that to happen as well and then slowly get more and more a high percentage of it. Automate it and just kind of going from there so yeah. See the way your portfolio responds to the model and you just kind of [inaudible 00:53:07]
Nic Ryan: You can go for something more complicated like a rainforest, you can maybe try some deep learning methods. To me, I tend to try the simple one and if that’s good enough I’ll stick with it. But generally speaking, it is often the big thing as well is with natural language processing for an application like that there’s even some very simple things that you can do by having just even a keyword look up for some of the fields that you know are going to be there as well, so it’s a combination of a few things as well. Attacking it from a few angles as well. So that’s bit of a trick as well.
Kirill Eremenko: Okay.
Nic Ryan: If you know for instance from a bank statement, there is a certain company that it always goes to, like in Australia, Woolworths for instance. Maybe that’s not local bunk to go to Woolworths Petrol. but that’s a shame that we don’t have Woolworths Petrol. That’s where the bi-gram’s coming but we just have Woolworths the shopping center. Sorry, the supermarket. And we always know that that’s going to match the groceries and so you just have a keyword for that, so that’s a combination of simple approaches I think is often good.
Kirill Eremenko: Okay. Very cool. Alright so natural language processing is a really cool thing to learn. Actually I was using … my brother has this thing, it’s called Google Home. Have you used it before?
Nic Ryan: Yeah, I stayed at my friend’s place actually in Brisbane and he was turning the lights on and off using Google Home. It’s great.
Kirill Eremenko: Yeah, it’s crazy. You can play something on your TV, you can switch the music on … it’s just such a cool tool and I find it’s actually … but the way, not affiliated with Google or anything but I find Google Home is better than Alexa from Amazon. I’ve used both and somehow Google Home just catches all the things you say, like 90 percent of the things you say. Very very cool. So it’s coming into our lives and great advice for anybody out there to start learning natural language processing as scary as it might be, it is going to be more and more widespread and experts in that field are going to be needed more and more as well. For sure.
Nic Ryan: Yeah I think so. Yeah, I think natural language processing is going to be good. But I think as well there’s going to be some pretty cool things happening with, as you say, audio and computer vision as well. So I’m starting to see some more of those projects coming through as well where people are looking at plans for houses and trying to work out cost of buildings and all sorts of cool things like that.
Kirill Eremenko: Gotcha. Okay, so we’re coming slowly to the end of the podcast. I wanted to ask you … you mentioned you liked mentoring and helping people progress whether its in basketball, it’s in their professional lives, in data science, and so on. What would your one biggest piece of advice be for someone who is starting out into the field of data science and they might have prior experience in other areas, maybe not, but they’re quite new to the space of data science, what would you say to them and what would your one biggest piece of advice be for them to help them become successful?
Nic Ryan: I think, for me, I was pretty late picking up weightlifting. I do gym 3 times a week and I’ve only really started that for about 2 years. And when you’re doing it, when you first start, it’s extremely painful. I was walking like Robo-cop that day and it was agony, and all I was doing was hurting myself and I didn’t see results and I wasn’t getting any stronger and it was just … so you sort of keep it up and you keep it consistent and it’s 3 days a week and you lock yourself into this routine. You keep doing it. And then, over time you start to slowly see results and it starts to slowly get easier and you’re starting to be able to do more and more.
Nic Ryan: And so, that’s a great analogy as well for learning data science. It really does … initially it’s painful, but if you stick with it and if you do set up a schedule you will have success every time and you will look back on it in a year and think “wow I know all this stuff” but when you’re sitting in the hurt locker doing it, it doesn’t really feel like it’s much fun at times. But that’s all I’d say is consistency is the key. And just making sure that you just keep going with it, and looking for again, whatever’s out there that can help you. And what you guys charge for what is really a quality education. You couldn’t go to a university and get the education that you guys are charging for a couple bucks. It’s just incredible. So there’s some great resources out there, so speak to people, find out what you need to know and get out there and do it.
Kirill Eremenko: Fantastic. Thank you. Very good metaphor. Totally loved it with the gym and yeah, I totally agree. Consistency is key. Alright, so before … first of all, thank you so much for coming on the show and helping us see this whole world of data science working remotely and being a consultant to other companies and combining the technical side of things and the data science advise and strategy side of things. Before I let you go, I would like to make sure our listeners can get in touch. What are the best places to find you, connect with you, and maybe follow your career?
Nic Ryan: Yeah, definitely. LinkedIn is probably the place to go. I’d say yeah, just connect with me on LinkedIn.
Kirill Eremenko: Awesome. Okay, cool so we’ll share Nic’s LinkedIn in the show notes. And I have one final question for you Nic.
Nic Ryan: Yeah.
Kirill Eremenko: What is a book that you can recommend to our listeners to help them succeed in their careers?
Nic Ryan: There’s a book that I have here that is really quite good. It’s The Structural Interpretation of Computer Programs and it teaches you how to code through scheme, and it’s more about methodology and understanding problems and decomposing problems and it’s actually a good book, but it’s a bit heavy. But what I’ve discovered fairly recently is a book called How to Design Programs that also uses scheme as a teaching language and I think for someone starting out … ’cause I think a lot of people go down the path of learning a language and not necessarily how to solve problems and this book is a bit different. And it also has supporting software as well [inaudible 01:00:21] that helps you out to build these programs and schemes so I’m a real fan of that book and I reckon that learning the methodology as opposed to a language would be good for someone staring out. And I wish I would’ve seen that book earlier. Hopefully someone finds that interesting.
Kirill Eremenko: That’s cool. So what’s it called again?
Nic Ryan: How to Design Programs.
Kirill Eremenko: Oh.
Nic Ryan: I’ll send you a link for it. It’s got supporting software. You can write your own little programs in scheme. It’s a functional programming.
Kirill Eremenko: Okay, gotcha.
Nic Ryan: It’s cool. It actually sort of teaches you to break apart a complex problem into little parts and be able to code little functions and it’s … I’m just starting it now but it’s really quite cool.
Kirill Eremenko: Okay, cool. So we’ll share that in the show notes. And on that note, once again, thank you so much Nic for coming on the show. Been a great pleasure chatting with you and yeah, hope to catch up sometime soon. Otherwise good luck with all your projects that are in your pipeline.
Nic Ryan: Oh, likewise Kirill. Absolutely a pleasure and thanks so much it’s been wonderful chatting to you.
Kirill Eremenko: So there you have it, ladies and gentlemen. That was Nic Ryan, Data Science Consultant from Bundaberg, Australia. My personal favorite part of this podcast was the whole combination of the two things that Nic did. I didn’t realize this before the podcast but it’s really cool to see that on one hand he does technical projects and helps companies actually with the code and the modeling side of things, and on the other hand he has a space in his career where he goes into companies and helps them with the data strategy, how they can apply data in projects, and also mentors their employees and staff. I think that’s a really cool way to combine those two aspects in a career and maybe that will be helpful to some of our listeners as well.
Kirill Eremenko: And as usual, you can get all the show notes for this episode at www.www.superdatascience.com/235. That’s www.superdatascience.com/235. All of the things that we’ve mentioned on this podcast will be there, including a URL to Nic’s LinkedIn. Make sure to hit him up and connect with him there as well. On that note, thank you so much for being here today. I look forward to seeing you back here next time. And until then, happy analyzing.
