Jon Krohn: 00:00:00
This is episode number 471 with Kirill Eremenko, the Founder and CEO of SuperDataScience.
Jon Krohn: 00:00:12
Welcome to the SuperDataScience podcast. My name is Jon Krohn, a chief data scientist and bestselling author on deep learning. Each week, we bring you inspiring people and ideas to help you build a successful career in data science. Thanks for being here today, and now let’s make the complex, simple.
Jon Krohn: 00:00:42
Welcome back to the SuperDataScience podcast. I’m your host, Jon Krohn, and today we have a most special guest, who will no doubt be well-known to many of you. The one, the only, Kirill Eremenko. Kirill created this very SuperDataScience podcast and was host of the show from its first episode in 2016, through to episode 431 on December 31st, 2020. That’s two episodes per week for over four years. A year prior to launching the SuperDataScience podcast, Kirill founded the SuperDataScience company, and he is the firm’s CEO today.
Jon Krohn: 00:01:22
Superdatascience.com, the namesake of this podcast, is a comprehensive online education platform for data science and related data specializations. Through the SuperDataScience platform and his Udemy courses, Kirill has taught well over a million students worldwide, helping to skyrocket countless data science careers. In today’s episode, we’ll catch up with Kirill a bit on what he’s been up to since I took over as host of the podcast in January. In particular, we’ll be focused on the curriculum Kirill has been cooking up for you, that allows people to land their first data science job in 99 days. We’ll describe all of the content packed into the 99 day curriculum, as well as options for slowing the curriculum down, if you like, and completing it in longer timeframes. Kirill will also detail for us, the five myths underpinning people’s misguided beliefs that a data science career is unobtainable for them and the five items you need to land a data science job, even if you have no prior work experience in the field.
Jon Krohn: 00:02:28
As the content I just ran through suggests, today’s episode will appeal primarily to folks who aren’t yet data scientists, but are interested in becoming a data scientist or a related data professional, such as a data analyst, a software engineer, an AI expert, or a business intelligence expert. More broadly, the episode will be of interest to anyone who would like to hear what Kirill’s been up to the past few months, as well as anyone who loves learning about the psychology of big career decisions, something that impacts each and every one of us. All right, you ready for this momentous occasion? Let’s welcome Kirill back to his podcast.
Jon Krohn: 00:03:13
All right, everyone. We have a very special guest here. What’s his name again? Oh, yes, Kirill Eremenko, the host of the SuperDataScience podcast. Welcome back back to your show, Kirill.
Kirill Eremenko: 00:03:26
Oh, thank you. Thank you, Jon. It’s great to be here. Well, it is now your show and thank you for having me. It’s a strange experience, feeling of being here as a guest, but it’s fun.
Jon Krohn: 00:03:40
Yeah, no doubt. And we have a great episode lined up for today. I can’t wait to talk about the topics coming up, 99 days to being a data scientist, but first I’m sure the audience would love to hear what’s going on in your life. Where in the world are you?
Kirill Eremenko: 00:03:54
I’m in Gold Coast, Australia. It’s a city about 100 kilometers south of Brisbane, quite a big city, it’s got lots of suburbs, I’m in one of them. And yeah, just spending time here with family.
Jon Krohn: 00:04:08
That sounds really nice. Especially with the pandemic and stuff, to be able to be near your family, that’s a nice thing to be able to do.
Kirill Eremenko: 00:04:16
Yeah.
Jon Krohn: 00:04:17
So, what have you been up to with all the spare time that you have, now that you’re not hosting the SuperDataScience show?
Kirill Eremenko: 00:04:22
That’s a good question. I was thinking about that earlier. Nothing really huge, just trying to run the businesses and help our management teams to improve them and help people around the world. So, nothing really specific stands out. Maybe I’ve been dedicating a bit more time to trying to improve myself and connect with parts of myself that I have neglected before, like the emotional side of me, the less analytical structured part of my brain or my being. And it’s been really tough, it’s been a really tough journey to unlock those parts of myself.
Jon Krohn: 00:05:20
Super interesting. Maybe we can dig into that at some point in the episode or it might come up organically.
Kirill Eremenko: 00:05:26
Maybe. Maybe.
Jon Krohn: 00:05:27
So, was it along this journey that you came up with the idea that a data scientist could be formed in 99 days alone? I think that this is a really revolutionary idea and I’m interested in how it even came about. So, as soon as you explained to me that you were doing this talk at DataScienceGO, on 99 days to landing your first data science job.
Kirill Eremenko: 00:05:51
That was at ODSC.
Jon Krohn: 00:05:52
Oh, was it ODSC?
Kirill Eremenko: 00:05:54
Yeah. Yeah. The same conference you were talking at.
Jon Krohn: 00:05:56
Oh, wow.
Kirill Eremenko: 00:05:57
Yeah. Haven’t done this talk at DataScienceGO yet.
Jon Krohn: 00:06:00
Oh, man, I just assumed, for sure. Oh, that’s even cooler. At ODSC Europe, ODSC East?
Kirill Eremenko: 00:06:06
The same one you were at, East.
Jon Krohn: 00:06:07
ODSC East.
Kirill Eremenko: 00:06:07
Yeah.
Jon Krohn: 00:06:10
Yeah. Oh, cool. Wow. I don’t know how I missed that specific detail.
Kirill Eremenko: 00:06:16
I usually talk at DataScienceGO, I guess that’s why.
Jon Krohn: 00:06:19
Yeah, I guess I just assumed.
Kirill Eremenko: 00:06:20
Yeah.
Jon Krohn: 00:06:21
But regardless where you did the talk, you did a talk on becoming a data scientist in 99 days, and the SuperDataScience platform now has a track that allows people to follow a curated curriculum, that allows them through specific coursework, specific projects, to become a data scientist in as little as 99 days. And I think I was about to say this a few minutes ago, that it’s a revolutionary idea and as soon as you mentioned it to me, it was so obvious because I do believe that it’s true, but I would often think, “Oh, you’re going to need months or years.” But in a lot of cases you could, if you dedicate yourself, become a data scientist in 99 days. So, how did it come about?
Kirill Eremenko: 00:07:10
So, randomly, by accident, I needed to prepare a talk for ODSC, because they reached out and said, “Can you come do a talk?” And I was like, “What should I talk about?” And they gave a few ideas for topics, and the things that related to me the most and that I loved the most were about careers and helping people either switch careers or get started in data science. And one of them was how to get your first job in data science, and I thought, “Oh, okay, how can we make this more fun and challenging and what can we do about this?”
Kirill Eremenko: 00:07:48
And then at the time, our team at SuperDataScience, so we’ve got a product architect and a data analyst who are working together to create different learning paths, and they were working on this learning path. And it was my job to review the courses that they put into the learning path, the sequence, the extra materials. And it just happened, these two things were happening at the same time. And by looking at it, I was like, “Oh, this is very interesting.” Because they had days, the number of days they thought that a person would need to complete this. And I don’t remember exactly, but in their interpretation it was six or seven weeks. Was it six or seven? Seven times seven? No, no. So, 17 weeks or 15 weeks or something … Four or five months to complete this.
Kirill Eremenko: 00:08:44
And I was like, “Oh, that’s really cool.” So, I went through that and I saw okay, for instance, this whole section is not really relevant on specific type of business analytics. It’s more of an application of data science rather than a fundamental data science type that you need to know. So, I thought, “Okay. So, if we throw that out, this can adjust and so on.” And it came down to like 115 days and I thought, “That’s not click-baity enough. It’s got to be something really cool.” And I thought, “Okay, how can we restructure this a bit more?” And I thought that, “Okay, if somebody really pushes themselves…”
Kirill Eremenko: 00:09:24
So the 99 days, it’s not for everyone. If you’ll imagine the bell curve, it’s for the people that really push themselves, the people one standard deviation away from the mean. You can get a data science job following this path probably in three months, six months. So, three months is 99, is like six months, maybe a year, but the more you push yourself, the faster. And I thought, okay in 99 days, if you do the first week or so …
Kirill Eremenko: 00:09:52
Actually I have it somewhere here, the first couple of days or a week, you do … Let me check it here. So, you do an introduction to data science, your first week, get an overview of some of the topics. And then, you do the basics of visualization for two weeks, then you do coding for data analysis for two weeks. You do SQL and databases for two weeks. You do applied business analytics, that’s more like statistics and data science, for a week. You’re doing machine learning for a week. So, that’s your core. That’s already your nine weeks. Then you have a catch-up week for week 10, and then weeks 11, 12, and 13, you specialize, you’re either machine learning, data visualization, or data wrangling. And then, you have another catch-up week just to have a buffer to catch up. And then, that brings it to 14 weeks. 14 times seven is what, 90, 98, right? So, 28 plus 70, 98. And so, the last day, 99, is celebration day because that’s when you land the job.
Jon Krohn: 00:10:52
Oh, nice.
Kirill Eremenko: 00:10:54
And the interesting thing about this is you not only go through this and learn the content and your data science, but in parallel, you got to be doing software career, networking, putting content out there, reaching out to people, reaching out to recruiters. So, if you do those in parallel, we believe, and this is just for everyone out there, the caveat is that this is a hypothesis, right? We have a hypothesis that it can be done in 99 days.
Kirill Eremenko: 00:11:25
We have success stories of students who have gotten jobs in data science based on our courses and other things, but we don’t have somebody who’s done it in specifically 99 days. We’ve launched this challenge and we have a cohort of people who are participating in it. So, there’s about just under 100 people, I think around 80 people who’ve taken the challenge and about 16 people who we’re actually tracking how they’re going through the challenge. So, that by the end of it, by the end of this first cohort, we have a success story or some success stories, proof that it can be done in 99 days.
Kirill Eremenko: 00:12:01
But if anybody really wants to participate, the challenge is open, it’s free to take part in. You can find it at www.superdatascience.com/challenge. And there you get … What you actually get with this challenge is, you get a massive, I think it’s like 30 page learning guide. Step-by-step what you need to do every single week, and if you follow it and if you get a job in 99 days, of course share that on social media, that you’re participating. I think the hashtag is #SDS99, but otherwise just tell us, email support@superdatascience.com and we’d love to share your story.
Jon Krohn: 00:12:40
That sounds really cool. So, do you imagine this is something that people can do while they have a job or is this something you need to do full-time?
Kirill Eremenko: 00:12:48
The way it’s structured is, it’s only a few, it’s I think, one or two … And I don’t want to be misinterpreting here, but I think it’s one or two or three hours a day. So, if you’re going to fit that in with your job, especially with working from home and maybe if you’re going to be at the same job, if you already have a job and you’re going to be at the same company, maybe the company would be happy for you to spend the time or if not … If you really want something … That’s the reason I was saying it’s not for everyone, it’s for people who will dedicate themselves a lot. If you really, really want something, you’re going to find a way. You’re going to wake up an hour earlier and go to sleep an hour later. Some way. Or do it during your lunchtime. There has to be a way to make it happen.
Jon Krohn: 00:13:34
Speaking from personal experience, it’s hard to imagine a career that you could be more excited about and want to pursue with that level of passion. I’m not being sarcastic, data science is such an exciting job, full of new challenges and new things to learn every day. And yeah, it wouldn’t be surprising to me if lots of people can even see that from a distance and say, “I can tell that, that’s going to be incredible,” and be passionate about pursuing this for 99 days.
Kirill Eremenko: 00:14:03
Absolutely, absolutely. It’s totally worth it. And then of course, it has to be something that you want, but we can assume it’s something you want if you’re listening to this podcast.
Jon Krohn: 00:14:13
All right. This episode is brought to you by SuperDataScience. Yes, our online membership platform for transitioning into data science and the namesake of the podcast itself. In the SuperDataScience platform we recently launched our new 99 Day Data Scientist Study Plan, a cheat sheet with week-by-week instructions to get you started as a data scientist in as few as 15 weeks. Each week, you complete tasks in four categories. The first is super data science courses to become familiar with the technical foundations of data science. The second is hands-on projects to fill up your portfolio and showcase your knowledge in your job applications. The third is a career toolkit with actions to help you stand out in your job hunting. And the fourth is additional curated resources, such as articles, books, and podcasts to expand your learning and stay up to date.
Jon Krohn: 00:15:09
To devise this curriculum, we sat down with some of the best data scientists as well as many of our most successful students, and came up with the ideal 99 Day Data Scientist Study Plan to teach you everything you need to succeed, so you can skip the planning and simply focus on learning. We believe the program can be completed in 99 days and we challenge you to do it. Are you ready? Go to www.superdatascience.com/challenge. Download the 99 Day Study Plan and use it with your SuperDataScience subscription to get started as a data scientist in under 100 days. And now, let’s get back to this amazing episode.
Jon Krohn: 00:15:48
So, perfect segue here. I hear that you have a list of five reasons why people might think that pursuing a data science career is impossible. So, the kinds of things that would come up in people’s minds that they think, “No, I couldn’t do this in a year. I couldn’t do it in two years. I couldn’t do to 99 days.” Yeah. You’re the master. What are these things?
Kirill Eremenko: 00:16:10
I wonder how you know I have this list, Jon? Yeah. A lot of things happened behind the scenes, before the podcast starts, we had an hour chat before we started.
Jon Krohn: 00:16:22
I installed spyware on your computer.
Kirill Eremenko: 00:16:27
There’s nothing to spy on.
Jon Krohn: 00:16:29
It’s really, really boring. You just create data science videos all day.
Kirill Eremenko: 00:16:34
Yeah. Oh, that’s the one thing I have been doing more of, I’ve been doing more of webinars. I think this year I’ve done the talk and I think I’ve done two webinars and I’m planning on doing a third one soon. So, I enjoy that. It’s always hard to get started. It’s putting together a good webinar for people to follow along and enjoy and also, first webinar I did this year, it’s supposed to be one and a half hours, went for three.
Jon Krohn: 00:17:01
Oh, that always-
Kirill Eremenko: 00:17:02
I’m like, oh, that is too much. I was tired at the end of it. Everybody was tired.
Jon Krohn: 00:17:05
Yeah. That always happens to me.
Kirill Eremenko: 00:17:07
Yeah. So, it’s like an arch to do it, but then once you get started, it’s so much fun and it’s enjoyable seeing everybody tune in and listen in.
Jon Krohn: 00:17:20
Quickly, before we get to the five myths, what were the other webinars on?
Kirill Eremenko: 00:17:24
Mostly on Tableau. So, one was how to land your first job in Tableau in 50 days.
Jon Krohn: 00:17:34
Oh, my goodness.
Kirill Eremenko: 00:17:40
Yeah. Because the reason why it’s shorter is because … That was also a good template for me for this talk, because I thought, “All right, if somebody can land a job in Tableau in 50 days, based on that webinar I prepared…” It’s also to do with certification, we have this course where in 30 days you can get certified with Tableau, and certification really helps you land a job because it’s like it gives you recognition. And I thought, “Okay, if somebody can land a job in 50 days with Tableau, which is a visualization tool, how long would it take in data science?” And the difference is, that somebody studying Tableau already knows about data science, already has done the fundamentals and understands that machine learning is probably not for them, data preparation or data wrangling is not for them.
Kirill Eremenko: 00:18:25
All this other stuff is not … Maybe it’s not what interests them the most, but data visualization is what they want. So, they already have a headstart in terms of the overview of data science, understanding what they want, because a big part of this challenge is, data science is big, you can’t specialize in everything. You have to choose a specialization pathway. So, the first couple of weeks are helping you understand, get a feel for all these different parts of data science and understanding what you want. And that’s why I reckon like, “Okay, if we can help somebody land a job in Tableau in 50 days, data science is going to take a bit longer, I think we can get it done 99.”
Jon Krohn: 00:19:00
Cool.
Kirill Eremenko: 00:19:01
Yeah, that was one of them. The other one was also on Tableau, but this one was shorter, and I think it was myths. Myths why people can’t get into Tableau or something along those lines.
Jon Krohn: 00:19:17
Are those on YouTube? Can people watch these?
Kirill Eremenko: 00:19:21
I’m not sure. Maybe, no, I don’t think those are on YouTube, unfortunately. Yeah. The team needed them for something else.
Jon Krohn: 00:19:30
All right. Well, let’s jump into these myths. We’ve talked about them, I’d love to hear more about them. The five myths that make people think data science is impossible to get into today.
Kirill Eremenko: 00:19:42
Okay, awesome. So, career limiting myth number one is, “I’ve missed the train, data science is no longer sexy”. I remember that article 2012, Tom Davenport, D.J. Patil, Data Science, Sexiest Job of the 21st Century.
Jon Krohn: 00:19:55
Yeah, yeah, yeah.
Kirill Eremenko: 00:19:57
I think data science was coined in 2010 or somewhere around there. 2012, that article came out. So, things started going really well for data science, especially after that. And then 2015, ’17, that’s when data science really peaked. That’s where we saw the highest enrollments ever into our courses on Udemy, highest interest, everybody was like, data science is the hot thing. We’re in 2021 now and maybe this has tempered off, tapered off, maybe this is no longer the thing, because it’s been over 10 years since data science started as a profession. It’s been almost 10 years since the article. It’s been five or four years since the peak of the popularity of data science.
Kirill Eremenko: 00:20:42
So, maybe it’s tapered off and it’s going to die off completely. So, that’s the first thing you got to answer for yourself before getting into this profession, because there’s a saying, “A rising tide lifts all boats, a falling tide makes all boats fall.” So even if somewhere in a specific company, data science is on the rise, if overall, in the whole of the world, data science is dying away, what’s the point?
Jon Krohn: 00:21:09
Yeah, yeah, yeah.
Kirill Eremenko: 00:21:11
Right. So, what I did to disprove that myth, I looked at the the Glassdoor, Best Jobs in America, and Glassdoor is a website that posts careers and you can post reviews of your career and so on. And so, they have this list of Best Jobs in America, it actually started in 2016 and they measure three things. So, median base salary, job satisfaction, and job openings. I think in 2016, the job satisfaction was something else, some other metric, but then they changed it to job satisfaction.
Kirill Eremenko: 00:21:49
And check this out, so in 2016, it was the number one job in the whole of America, data scientist. 2017, data scientist. No surprise, right? Because it was the peak of popularity. 2018, data scientist. 2019, data scientist. 2020, data scientist drops to third place, after front-end engineer and Java developer. And 2021 … So, this is data for the previous year obviously. It comes out at the start of 2021 and they use data for the past year. So in 2021, data scientist went back up to the second place.
Jon Krohn: 00:22:21
There you go.
Kirill Eremenko: 00:22:22
So, it’s been in the top three for the past, what six, seven years? That’s already good evidence. And then, just to top it up, I looked at a LinkedIn report called The 2020 Emerging Jobs Report. Where better to look than a place where everybody posts about their jobs? So, this Emerging Jobs Report, what they did is the methodology of how they created it is, they took people who’ve been in the workforce for the past five years. So, to qualify for this analysis, they basically just trawl through the data. They’re not going to ask you, they’re just looking through their data. But they only pick people who’ve had a full-time job for the past five years, and they looked at the changes of titles, how many people changed their title from X to data scientist? How many people change the title from X to AI specialist? And so on, or whatever, or I don’t know, auditor or accountant and so on.
Kirill Eremenko: 00:23:23
That’s how I understand it, please check the 2020 Emerging Jobs Report by LinkedIn to make sure my interpretation is correct. The top job in 2020, according to this methodology was artificial intelligence specialist. It grew by 74%. So, 74% more people got the AI title compared to people before the start of the year.
Jon Krohn: 00:23:51
I mean, that’s a data scientist.
Kirill Eremenko: 00:23:53
Yeah, that is a type of data scientist, funnily. Second one is robotics engineer, 40%. And the funny thing about this is, I looked at the report and you will see a photo on the report, a stock photo of two people standing in front of a robot and they’re trying to adjust it or understand it, on this table. And the one’s like this, propping his head, the other one’s showing him something. And then, I look at the descriptions and it’s like, process automation, UI path, automation anywhere. All this robotics process automation software, not robotics engineering.
Kirill Eremenko: 00:24:34
So, I think they got the picture wrong and I think they also got the title wrong because the description, except for one word or one subtitle in that description, the rest of them are all to do with RPA, like automating high volume, rudimentary tasks in a business, nothing to do with physical robots.
Jon Krohn: 00:24:58
And that’s a data scientist.
Kirill Eremenko: 00:24:59
That’s also a data scientist. In my view, RPA is a type of AI, very basic, trivial type of AI, but it’s still got to do a lot with data, understanding processes and automating them. And then, finally in third place, data scientist, with a 37% growth rate. So, whichever way you look at it, either one of the top three is data scientists or are all of the top three are data scientists, but this profession is on the rise.
Jon Krohn: 00:25:24
The other key thing there is, for sure, that third category, the actual data scientist one, that one has a much bigger starting base than the other two, for sure. I know, from knowing the market, that those first two AI specialist and RPA specialist, it’s a much smaller … There’s probably at least 10 times as many people, maybe it could be 100 times as many people with the title of data scientist relative to those first two. So, even though percentage wise, it isn’t as big of a growth, in terms of the total volume it’d be huge.
Kirill Eremenko: 00:26:01
Yeah, absolutely. That’s a very good point. I think that’s enough stats to consider this myth busted.
Jon Krohn: 00:26:09
Well, it’s busted for sure. And I mean, there’s lots of other things. I was just reading yesterday that there was over 400,000 open data scientist jobs in the us. Yeah.
Kirill Eremenko: 00:26:20
Crazy.
Jon Krohn: 00:26:22
Definitely still sexy.
Kirill Eremenko: 00:26:25
It’s still sexy.
Jon Krohn: 00:26:27
What’s number two?
Kirill Eremenko: 00:26:28
Number two is a common myth or a common belief that “I can’t get a job because I don’t have any experience”. Have you heard this myth? [crosstalk 00:26:40].
Jon Krohn: 00:26:39
Yeah. I mean, it’s a classic circular chicken egg job issue.
Kirill Eremenko: 00:26:44
Yeah.
Jon Krohn: 00:26:45
But yeah, you definitely hear this. And I’ve had people… I’ve had students who have taken the courses of mine in-person or online who say, “I’ve been looking for two years to get into data science, but no one will hire me as a data scientist because I don’t have data scientist on my resume.” But I mean, I already know this is a myth, but I’m not going to spoil it, so please tell us. Please, go ahead.
Kirill Eremenko: 00:27:07
Okay. So, the way I approached these next three myths was by looking at examples of people who’ve come on the podcast, on this podcast, previously. So, when I was hosting it, people I’ve interviewed. And I went through a lot of potential candidates for this. I knew I only had 30 minutes for this talk, so I couldn’t include everybody. And there were some really cool stories, but I had to pick just one per myth, and picked one that I guess resonated with me the most.
Kirill Eremenko: 00:27:49
And for this first myth, we’ve got the story of Nicholas Cepeda. If you haven’t heard his podcast, it’s episode 41, it’s called An Inspiring Journey from a Totally Different Background to Data Science. And Nick’s story is he started off… What did he do? So, he worked as a business intelligence intern just for three months in 2016. And prior to that, he was in the United States Marine Corps, where he served for four years in aviation supply. And he was, I think, running some basic SQL queries in terms to get numbers of stock or something like that. But in his own words, he said, “I don’t really have that analytic experience. And most of my experience comes from data science courses, college. Business intelligence courses are new to me.”
Kirill Eremenko: 00:28:46
So, no experience. Apart from the three months internship as a business intelligence intern, no experience in data science. And he goes and applies for a job at the Walt Disney company as a data analyst in marketing. And so, what does he have to show for it, right? He doesn’t have the experience, wouldn’t you just like turn him around if you were the hiring manager? But he’s taken quite a few courses on data science, so he took our course on R programming, advanced one. He took a big data and Hadoop development course, he took our Tableau advanced course, our Tableau 10 course, our machine learning A-Z course, and a course from Dell on data science and big data analytics.
Kirill Eremenko: 00:29:34
And so, when he rocks up to this interview, and you can hear him talk about it on the podcast, they ask him to talk about like… Okay, I don’t remember exactly what the question was, but basically he ended up talking about a case study that we did in one of the Tableau courses on this company, this imaginative company that washes your laundry, picks up your laundry, goes away, washes it, brings it back to the next day, and how they needed to expand their operations in the US and they had these two markets, and he needed to do clustering. So, he did clustering in our course, but in Tableau, not even in Python or machine learning, but in Tableau.
Kirill Eremenko: 00:30:17
And he explained his experience from the course. He’s like, “Oh, so there was this company. It’s a mock company. But nevertheless, we needed to find what kind of three markets in the US that they hold. And there was nothing, no prescribed markets to look at, so we had to do clustering, and this is how it worked, we put these trendlines and so on.” And in his own words, he said, he mentioned that project, and it sounded like her eyes lit up. Like the hiring manager that was interviewing him was very excited because that exact thing could be applied to what they needed in the company.
Kirill Eremenko: 00:30:55
And then the second thing was they asked him about like, “Oh, can you deal with real world data? With messy data and so on?” Because that’s a good point. In courses, often you get sanitized nice and pretty data so you can just follow the steps and get everything working. We have a course on, I think, advanced R programming where we spent quite a bit of time on messy data preparation. That’s one of the focuses of several of our courses. And yeah, he told them that, so here’s a quote from him, he told them, “I told them that 70% of time is spent on creating strategy on how I’m going to prep and clean the data.” So, he already had that experience. That 70% of the time of a data scientist is data prep. And he told them about the median imputation methods, like averaging things or other approaches that you can take for missing, corrupt, messy data and how to deal with it.
Kirill Eremenko: 00:31:50
Again, something he just learned from a course, right? And as a result of an interview like this, that is based on his experience from case studies from courses, he landed the job. So, he’s been at Walt Disney company. It was funny, because during the podcast he couldn’t say the name of the company because he hadn’t landed it yet. But just before it went out, he emailed me and said, “Hey, can you add this in the intro?” And I did add it in the intro because he landed the job. And he’s been there for three years, 10 months. He started in advanced analytics and optimization. And recently he… well, not recently, that was five months, then he was a senior marketing insights analyst. And I think he actually maybe even went higher in his career at Walt Disney recently.
Kirill Eremenko: 00:32:38
The moral of this, I guess, is that recruiters, hiring managers, they don’t really care about your work experience. What they care about is your capacity to deliver on the projects that they have. And work experience is a proxy for evidence of that capacity. So, if you have work experience, like three, five years, they can kind of say, “Oh yeah, well, that person probably knows what they’re doing, probably can deliver. They have the capacity to deliver.” But if you can illustrate this capacity in whatever other way, as a hiring manager, why would I care? Why would I care if you have zero years of experience or 50 years of experience, if that’s even possible, if you can deliver on the needs that my business has? You just need to be able to demonstrate that, and thereby work experience not really an absolute requirement.
Jon Krohn: 00:33:34
100%. And yeah, I thought you might go down the route of saying, and this is related, is maybe your first job in the field isn’t going to be with a title of data scientist. And that’s another like… So, this person was already working for a few months in business intelligence, so that can be a stepping stone to a data science career.
Kirill Eremenko: 00:33:57
Yeah. And we’ll talk about that in the fifth myth. We’ll get there.
Jon Krohn: 00:34:02
Nice. All right. Well, what’s the third myth?
Kirill Eremenko: 00:34:05
Okay. Third myth is “I have the wrong background”. Heard that one before?
Jon Krohn: 00:34:10
Oh yeah, for sure. Then people will-
Kirill Eremenko: 00:34:12
I studied arts, or I studied psychology, or whatever else.
Jon Krohn: 00:34:17
Totally.
Kirill Eremenko: 00:34:17
How can I get into data science?
Jon Krohn: 00:34:19
Yeah.
Kirill Eremenko: 00:34:19
Yeah. So here, again, I get plenty of examples on the podcast of people who’ve come into data science specifically from those things, from psychology, I talked to somebody in Switzerland who studied psychology then came into data science. And San Diego, I met a lady who was in some arts background and I think writing, or maybe even theater or something like that. And she wanted to get into data science and we came up with some ideas of what value she can bring to the field. And this specific example I used in this myth is the story about Ayodele Odubela.
Jon Krohn: 00:35:00
Oh, no way.
Kirill Eremenko: 00:35:02
You met her, right?
Jon Krohn: 00:35:02
Yeah. She was on an episode really recently. I’m looking up the number right now, but you can talk and I’ll mention it [inaudible 00:35:09].
Kirill Eremenko: 00:35:09
Yeah. So, the first time she was on the podcast was episode 297, and the title is Fortitude & Passion in the Data Science Journey. And Ayodele’s story is super interesting because she actually studied, I don’t know if you talked about this with her in the episode that you recorded, but she studied an associate of arts in radio, TV, and film. And then after that, she went and did a bachelor’s, a bachelor of arts again, in media and professional communications. And check this out, these are the titles of some of… I went onto the University of Pittsburgh website to just look at, “Okay, what do people study in this program? In this course?” Here’s some of the titles from the course, these are subjects that people study. Organizational psychology, audio storytelling, photography, communication, fundamentals of filmmaking. How much further away from data science can it get?
Jon Krohn: 00:36:12
Yeah, except for obviously the need to be able to communicate data effectively, which is a hugely valuable skill. Which sometimes people think of it as an add-on, but it’s a critical thing to being… And in fact, a very recent episode, so first Ayodele was most recently on the show for episode 449, and we primarily talked about AI ethics. So, we didn’t talk too much about her background, but we talked about ethical issues with machine learning software. And something that totally blew my mind that I hadn’t thought about at all up until that episode, but have now read a lot about, is how hardware that collects data is also often very biased. Anyway, so it was an episode on AI ethics.
Kirill Eremenko: 00:36:59
Whoa.
Jon Krohn: 00:36:59
Yeah. So for example, I was reading just this past weekend. I was reading in The Economist how… I can’t remember what the… Oh, pulse oximeter is the name of the device. So, it’s a device that measures your blood pressure. You just wear it on your fingertip. And it was designed-
Kirill Eremenko: 00:37:17
The one that they use in hospitals?
Jon Krohn: 00:37:19
They use in hospitals. And so it’s to quickly get you a sense of, “Okay, how oxygenated is the blood?” And so for example, in COVID, having below normal blood oxygenation levels, it could indicate that you need to stay in the hospital as opposed to going home. And these devices are designed by teams that consist almost entirely of white men. So, they don’t work as well on women because they don’t fit properly. So light gets in, and so you can get bad readings. And so, basically people who should be staying in hospital don’t stay in hospital. And then same thing, the darker your skin color is, there’s also higher error rates. And so, a African-American person with the same blood oxygenation level as a white person, the African-American person could get sent home, while the white person stays in hospital because of an erroneous reading. And so, this is something that we know has been happening through the COVID pandemic and is related to hardware issue. So anyway, it was a really interesting episode on AI ethics and I learned a ton. And actually, we selected that as the episode of the month for March.
Kirill Eremenko: 00:38:29
Wow.
Jon Krohn: 00:38:30
And then on the later point you were talking… So, the point of communication is, as I’m sure you also did a question that we often ask guests on the show is, what do you look for in people you hire? And there’s two things that come up all the time. The first one is communication, and the second one is the ability to learn. And basically, sometimes those are the only two that people mention. It’s like those are the things that you’re looking for in people you hire, because data science is such a quick moving field that those are really the key things.
Jon Krohn: 00:39:05
Anyway, so did a big episode on that recently, a Five-Minute Friday, episode 466 on Good vs. Great Data Scientists. We talked a lot about communication. So anyway, so all going back to Ayodele studying this course, which I’m sure is hugely… her background, maybe, yeah, so people I guess don’t think about it as a quantitative discipline and being directly relevant to data science. But in fact, all kinds of backgrounds can be hugely beneficial, like communication background and data science.
Kirill Eremenko: 00:39:39
Absolutely. And I’m glad you mentioned that point because I agree that whatever background you’re coming from, there’s something that you’ve learned, there’s certain things that you’re good at that will already benefit you. You might not even know what exactly it is, might be some business skills, it might be some accounting, it might be some communication, it might be some creativity even, it will come up sooner or later. So, even if it doesn’t help you get into the job at the start, eventually it’ll come up as a superpower, as like a secret weapon that you have and that you’re going to be using to your advantage.
Kirill Eremenko: 00:40:18
Okay. So, Ayodele’s story. In her own words when she was taking this course at University of Pittsburgh, there was one class on computer science and she actually… I think she stopped taking that or changed that track of her course because she didn’t like the computer science side of things. And then, after her uni degree, she got into, by way of chance of her profession, she got into marketing at first, but then she ended up doing a lot of data work in marketing. And she was a senior data analyst at YooLotto, basically performing analytics tasks there.
Kirill Eremenko: 00:41:07
And she mentioned that she leveraged SQL to analyze behaviors, she performed A/B tests, she saw where the market was moving, like the data science market, and decided to follow it, and she enjoyed it. And so, only after that, did she enroll in the master’s of data science. And since that first job as a senior data analyst, she’s had a job as a director of machine learning, a data scientist in another company, a part-time machine learning instructor on LinkedIn, and she’s now a data science evangelist. So, very interesting in the sense of how somebody from an arts background is able to land a job in data science. And then, not only that, but also skyrocket their career from there. And so, I think she’s like writing a book now, right? Or publishing a book?
Jon Krohn: 00:41:53
Yeah. So exactly to your point, talking about things like your communication education background allowing you to have a superpower once you’ve started to become a data scientist. Now that Ayodele is in that data science evangelist role, she is really taking advantage of that strong communication background. And you can tell from these episodes that she’s done, she’s an unbelievably clear communicator, she comes up with so many examples right off the cuff, it’s incredible.
Jon Krohn: 00:42:19
And so, she gets to be in this job now where she still has to be a data science expert, and she’s constantly learning new data science techniques. And she’s talking to prospective clients, prospective users, and showing them how they can be integrating those approaches into what they’re doing with common ML. Super cool job. And yes, exactly. So, she self-published this year a book on Getting Started in Data Science, if I remember correctly. And I don’t know what combination of things, but she has now landed a book deal with Wiley, one of the biggest publishers in the space, and that is a book on ethics and data science.
Kirill Eremenko: 00:43:01
Okay. Wow. Very cool. Such a story, right? Like going from, “I don’t really like the… what’s it called? The computer science side of things.” To, “I’m now publishing books on data science.” That’s incredible I think.
Jon Krohn: 00:43:15
Another cool thing is if listeners want to meet Ayodele, you can meet her on any Sunday that you want, because she hosts Office Hours with Harpreet Sahota, who was also on a recent episode, on episode 457. So, Harpreet and Ayodele cohost Office Hours on Sundays, you can find out about it on LinkedIn through their profiles. And so, they’re just Office Hours, you can just show up and ask career questions or data science questions or whatever. And they and other people will answer the questions. Cool thing.
Kirill Eremenko: 00:43:46
That’s really cool. A few people have attempted Office Hours, regular Office Hours and data science. I remember Randy Lao, I think, and Favio Vazquez, or maybe Kristen Kehrer and Favio Vazquez were doing Office Hours like a few years ago, maybe two or three years ago. And so, that is I think a very cool addition to the community. It’s hard though. It’s hard to keep this regular schedule, as you would know with the podcast, every week you need to record an episode. And hopefully Harpreet Sahota and Ayodele will be able to keep it on lock for as long as they can to help as many people. So, if you’re listening to this, I would jump on this and take advantage of it. Anything like that might not last very long. Hopefully it will, but might not. So, if you can now attend those Office Hours and see what value you can get, what questions you can ask.
Jon Krohn: 00:44:46
Yeah. I would say one of the biggest risks to it continuing isn’t their dedication to doing it, but is that their careers blow up so much that other opportunities come up.
Kirill Eremenko: 00:44:54
Yeah, yeah. That’s exactly the thing. So yeah, jump on the train while you can.
Jon Krohn: 00:45:00
Anyway, so yeah. So, that’s myth number three dispelled, not having the right background.
Kirill Eremenko: 00:45:04
Absolutely.
Jon Krohn: 00:45:05
What’s number four, Kirill?
Kirill Eremenko: 00:45:06
Okay. Number four, “online learning is not official”. I’m doing floating quotation marks. Is not official enough. And that’s a common myth about, “Oh, well, I need to go and either work in data science or do a master’s in data science or some other degree or something like that, in order to get my first job and in order to land my first job.” There’s nothing wrong with doing a master’s in data science, it’s just not a prerequisite for you to land a first job. You don’t need a PhD or any kind of real world in-person type of degree when there’s other ways of doing it. So example, Sean Casey, episode 383, You’re Not an Imposter, You’re Learning: Data Science Journeys. Very interesting because the way I met Sean was at one of our DataScienceGO events. I think it was our first DataScienceGO virtual, it was in June or July, 2020. And we have this feature, which you might know, where you click a button and you get to meet a random person from the world. Have you tried that?
Jon Krohn: 00:46:16
I haven’t. I feel bad that I haven’t. I feel like I should have done it.
Kirill Eremenko: 00:46:22
It’s okay. You should try it, it’s so much fun. I love doing it. I didn’t do it this time around because the event was like at 4:00 AM for me in Australia time. But whenever I can, I try it out. And you just click a button and you get to meet another person for five minutes, you chat, and then boom, another person from anywhere in the world. And so, I got connected with Sean Casey. I don’t remember at the time where he was, either Dubai, because he lived for a long time there, or he was already back in Ireland. And we just basically connected for… was supposed to be five minutes, but it was like 30 seconds before the connection cut out.
Kirill Eremenko: 00:47:02
And he started telling me his story. And so, then he later… I think I reached out to him on LinkedIn, because you can also get your LinkedIn details of each other. And then I asked him to send me a story by email, and he sent it to me, something like [inaudible 00:47:16]. I was like, “Wow, this is so cool.” So, I invited him to the podcast then. It was really interesting. So, his story goes as follows. He finished a bachelor degree in mathematics and computer science in 2005. So, this might be a weird start, right? We’re talking about not having that… online learning is not official enough. And the example we’re looking at is somebody who did a real world bachelor’s degree in computer science. So, what the heck are you talking about, Kirill? The thing is, that was back in 2005, so that was like years ago, long time ago.
Jon Krohn: 00:47:51
Be-Fortran.
Kirill Eremenko: 00:47:53
Fortran, yeah. And he himself told me that the things that later he needed to do in his data science career and that were required in the job application, not what he studied in that degree. And a lot of time had passed, it’s not relevant. And moreover to illustrate what he was doing afterwards that the time had passed, was after that he went and taught mathematics and became a mathematics computer science teacher at an educational system in the UAE. As I understand it was for a primary school or high school kids, not universities, like schools. Then he was another teacher at another school. Then he was a mathematics teacher at a higher colleges, another kind of school, I guess. So, he did that for five years. And then he was a senior educational consultant for eight years. So, in 2005 he does his degree in [inaudible 00:48:52] computer science, which is quite outdated. Then 13 years he spends in the education industry. And then somebody, his friend, tells him about data science and he’s like, “Oh, maybe I should check it out.”
Kirill Eremenko: 00:49:04
And he’s like, “How do I get into this field?” To your point, I have a full-time job in education, I can’t go and study data science master’s or anything. So, he signs up to courses online and he does… I have a slide here which we can’t share because this tool, it doesn’t have screen sharing, and you won’t be able to hear the audio anyway. But the point is, in this slide I have a screenshot of his LinkedIn, and he has taken four courses on Udemy about data science, five courses on edX, and eight courses on DataCamp, right? So, that’s 17 courses on data science, and these are big courses. Some of them, machine learning, Python, R programming, dashboards. Substantial investment of time. And the way he did it, he said he was driving to work in Dubai, he was listening to the course on his phone. You imagine listening to a Python programming, like what? And then on the way back he was listening to it again. And when he got home, he would redo that same tutorial, but now he had heard it he already knew what to look out for. And he did that.
Kirill Eremenko: 00:50:14
That’s what I call dedication, right? You find time driving to work to listen to a video course. Of course, he wasn’t watching the video, he was just listening to it. So, that you can do it more efficiently and get more value out of it for yourself when you’re at home. People who really want to get a career, find ways. Anyway, so he did that, and eventually he was able to, in 2018, land a job, a senior researcher insights and analytics. Then he progressed to manager of data and analytics in 2019. And now he is a data and analytics principal consultant of some division called Transformation Services. So yeah, totally possible. You don’t need to have a official degree just to… for full disclosure after this, after he’s landed a job, after he’s in the space of data science, then he decided to, “Okay, now I’m going to go do a master’s in data science.”
Kirill Eremenko: 00:51:23
But the point is, I think he’s still doing it now. I don’t know if he’s continuing it or not. But the point is, it’s not a prerequisite, right? With online learning, you can still get all the skills you need. And it goes back to the point that I mentioned before, that you just need evidence that you can get the job done. It doesn’t matter if it’s an online degree or real world degree.
Jon Krohn: 00:51:43
100%. And not only does it not matter whether you get the education formally at a university and get the degree, or do it online, from my perspective, the strongest way that you can present to me that you’re a data scientist is outside of those completely.
Jon Krohn: 00:52:03
So, you have a computer science background say, like Sean did, and you say, “Okay, I was building this web app as a hobby, or I was thinking about making this business or whatever, and then I realized I was going to need to have these data collected, or there was this opportunity to use machine learning in my web app. And so, I learned how to do it.” Even outside of the structured path of a university program or an online course, you’re going through stack overflow and GitHub repos and just figuring out how to do it and solving the problem that you were trying to do, without anybody putting you on a specific path. That kind of independence and that ability to have the courage, to believe that you could do something that you don’t know how to do, that is the most amazing thing when people can explain that to me in an interview.
Jon Krohn: 00:53:07
I think that’s what often makes somebody a really great data scientist. It’s typically something that people develop on the job after years of experience, but there’s no reason why you couldn’t without any formal education in anything quantitative or computer science or anything, you could come from an arts background and just be curious and figure out yourself how to create some data-driven application or model. And that, to me, that courage, that independence, is the most valuable thing.
Kirill Eremenko: 00:53:41
I love that point. And with your permission, I would like to take it even further and say, if you don’t have that, this might sound harsh and radical, I think it might be a good idea for you to consider leaving data science, leaving this field. And that goes for any field. If you want to become a master chef at a Michelin star restaurant, and you don’t have the curiosity to cook in your free time and to find new recipes and maybe learn a new style of cooking, you just want to be a chef, maybe you want it for the wrong reasons, right?
Kirill Eremenko: 00:54:20
If you really, truly want something with your heart, it’s going to be part of your nature, you’re going to be curious about it. It’s natural for humans to be curious about things that we care about. So, if it happens that somebody wants a job in data science, just because of the high salaries or just because that’s what their friends have, or just because that’s the new department their company’s building, if it hasn’t caught in your attention so much, it hasn’t caught in your nature and your being so much that you’re constantly curious about it, wondering what else you can learn and maybe reading a little bit about it.
Kirill Eremenko: 00:54:59
Doesn’t have to be huge investment of time and it can be different for everybody. Somebody might be into, “Okay, what kind of apps I can use for my own hobby websites?” Or somebody might be curious, “Oh, what kind of research has been done at Facebook or Google?” Or somebody might be curious about the top visualizations that are published on Tableau public. There has to be something that is catching you all the time that you’re constantly progressing in. And that’s a good indication that you’re on the right path. If you don’t have that, ask yourself why? Why don’t you have that? Is it because maybe you’re not allowing yourself the room for that or the time or the room for that output and for that curiosity? Or is it simply because you don’t really care, deep down you don’t really care that much about it and then you might just be in the wrong field.
Jon Krohn: 00:55:56
Strong point, but I don’t disagree with you.
Kirill Eremenko: 00:56:01
Yeah. Yeah, it’s hard to be sincere with yourself. That’s one thing I’ve been learning over the past four or five months, sincerity with yourself takes courage and it takes also setting aside your ego like, “Oh, I want to be paid a huge salary. Oh, data scientist is a great profession.” Or whatever else in life. You have to set aside your ego, you have to have the courage and you have to face your fears. But sincerity of yourself is the first step to being sincere with others. If you can’t be sincere with yourself, you are going to be insincere with others and who wants to live a life of lying? Lying to yourself, to others. Even though you’re convincing yourself it’s the truth, but deep down somewhere you know, part of you knows that it’s not the truth. So, seeking sincerity, I think, is one of the ultimate journeys in life.
Jon Krohn: 00:56:58
Yeah, beautifully said. All right, so getting us maybe a little bit back on track, have we finished number four?
Kirill Eremenko: 00:57:06
Yes, we have.
Jon Krohn: 00:57:07
Okay, nice.
Kirill Eremenko: 00:57:09
We have made the point. Myth number five won’t be about a story, so I just wanted to recap if you want to listen to any of those episodes, there are numbers 41 with Nic Cepeda, 297 with Ayodele Odubela, and 383 of Sean Casey. Myth number five, we kind of touched on it a little bit already, it’s about “landing a job in data science takes years”. The thing to keep in mind here is that a data scientist is a very unstructured profession at the moment. It’s not like accounting, where an auditor does this, a tax accountant does this, a financial accountant does this. An inventory accountant, if that’s a profession, does this. There’s certain types of accounting, there’s certain things you need to know, there’s a CPA exam that you need to pass and other things. It’s very structured.
Kirill Eremenko: 00:58:05
Whereas, in data science, somebody who’s called a data scientist in one company, might be called data analyst in another company. And then in a third company, they might be called an insights analyst. Somewhere else, they might call you something else. Moreover, a data scientist in company A, does machine learning. And data scientist in company B does data visualization and a data scientist in company C, does research. There’s no clear cut way of, well, what is a data scientist and what does this profession mean? So, don’t get caught up on going for the title, “I have to be a data scientist.” Look at the job description. What are you actually going to be doing? Maybe the title is business intelligence analyst, but you’re actually going to be doing what you want to be doing. Or it might be a graduate analyst or a junior data scientist or a machine learning engineer, AI specialist.
Kirill Eremenko: 00:58:56
So there’s lots of different titles and while it might be more challenging to get a data scientist role at your particular company that you’re interested in, maybe you can start in a different role, called an insights analyst. Basically, if you’re finding that it’s complex to get a specific type of job, open up your horizon, broaden your view and look at other options. For instance, my story, I was a graduate analyst when I joined Deloitte, then I was promoted at a data analyst. Then when I, two years later, went into the industry, into the financial industry, into a superannuation company, which is the word for pension in Australia, pension from a company, I started as an insights analyst there and even though I wanted to be a data scientist, I was like, “Oh wow, that would be cool.”
Kirill Eremenko: 00:59:47
But at the time I was like, “Okay, I will…” I think, at the time, I wasn’t even too fussed about data scientist, because I hadn’t explored it that much, so I started as an insights analyst, but then at some point I was like, “Okay, I want to be a data analyst.” What I did is, this wasn’t even a year into my role there, I went and I negotiated with my manager after, of course, proving that I can do the work, did some great projects and after that I went to negotiate. I didn’t negotiate for salary. I didn’t negotiate for benefits or other things. I just said, “Hey, I would like to change my title from insights analyst to data scientist and here’s why. These are the things I’ve been doing, you only have one insights analyst in the company, which is me and maybe we can change the name of the role, what’s that going to cost you? No big deal for your company.”
Kirill Eremenko: 01:00:41
They thought about it, they were like, “Yeah, sure.” So they officially changed my title to data scientist. Boom. There you go. I didn’t even have to apply for a data scientist job, just negotiation inside the company. So, every story is different and the point is, you want to get into the field of data science, you don’t necessarily want to land the data scientist job, because they might all be called differently.
Jon Krohn: 01:01:06
Perfectly said, I agree on everything. You have an open and shut case here, Kirill, on your five myths, so.
Kirill Eremenko: 01:01:16
Yeah, but I wanted to ask you? You’re a director of data science, right?
Jon Krohn: 01:01:23
Yeah.
Kirill Eremenko: 01:01:23
What do you guys call the most junior level of data scientists at Untapt?
Jon Krohn: 01:01:33 Yeah, so I’ve been the Chief Data Scientist at Untapt for six years and the most junior role… Well, I mean, I guess it depends on how you define it, right? So, I’ve hired two data scientists now, where data scientist was their first job. Actually, every single person I’ve hired as a data scientist, it was their first data science job.
Kirill Eremenko: 01:02:04
Wow, that’s awesome.
Jon Krohn: 01:02:07
Yeah, I hadn’t even-
Kirill Eremenko: 01:02:10
But, Jon, they don’t have any data science experience, why did you hire them?
Jon Krohn: 01:02:15
They were obviously awesome. And everyone has a different background, but it was obvious that they already were data scientists, that they deserved that title. So, one of them, he already had a 20 year career working as a software engineer at JP Morgan and also at a hedge fund and then he took a 12 week or 16 week data science bootcamp full-time and all of those things, that brief experience in the bootcamp where you’re thrown in, that’s kind of like your 99 days curriculum where you really throw people in the deep end and you have them go through these structured weeks, but you also have those 20 years of experience doing software development, which is highly relevant. We’re putting machine learning algorithms into production systems and so, he’s hugely valuable every day across our application in terms of systems architecture, backend engineering, but he wants to be a data scientist, so as much as possible, he’s training models and generating visualizations and he does more and more of that all the time.
Jon Krohn: 01:03:44
Another data scientist, he was working in hospitals as something called a neuro physiologist, so supporting surgeons when someone’s having live brain surgeries, so this is a highly technical person. He already had a master’s in biomedical engineering and he was, while doing this neurophysiology, supporting surgeons, which is also, you’re monitoring data collection hardware, you’re looking for patterns in data, there’s a data science element to that work and in real time, where someone’s life is on the line, so it’s serious data science in a way, right?
Jon Krohn: 01:04:27
And he was, in the evenings, taking the Georgia Tech master’s in computer science, specializing in AI, which is an extremely inexpensive, formal education option that you can do online. And also, because that was kind of more focused on the theory, he did lots of Udacity courses, getting hands-on experience with projects and that kind of thing and he did some volunteering at a company that was recycling fabric, so building computer vision systems to be able to recognize different recycled fabrics and he did that as a volunteer project to get a foot in the door as a data scientist.
Kirill Eremenko: 01:05:06
To your point of extra curiosity, extra interest, something outside your normal job.
Jon Krohn: 01:05:12
Yeah, exactly. So, there’s all these and then another one is a more traditional approach, where doing a PhD in quantitative discipline, but even then, making sure that you’re bolstering that formal education, getting a PhD in quantitative discipline, okay, that doesn’t necessarily mean that you’re actually ready to be a data scientist. So, on the side, he was doing his own projects, self starter projects, taking courses online, like Udemy courses, getting familiarity with a modern data science stack and doing his own projects, self-driven projects. And so, in all of these cases, so three examples, but three completely different examples of people getting their first data science job, getting that title from different kinds of backgrounds.
Kirill Eremenko: 01:06:03
Absolutely. Absolutely. Yeah, very interesting. And also goes out, there’s maybe managers and entrepreneurs listening to this, you can get really qualified, talented people to do data science work at your company, even if they don’t have data science experience, right? I think it’s a handicap that you’re putting on yourself if you’re only looking for people who’ve been in data science, who have data science experience. You need to also open up your horizons and look for people who are just passionate and who can get the job done.
Jon Krohn: 01:06:47
100%. I mean, I guess in a way, all words are made up, but data scientist is a recently made up word and it’s so nebulously defined. You gave a perfect example there of accountancy having very specific credentials that you need to have to do that or like nursing, you have very specific credentials, but data science, no such thing exists and the whole title, as you say, at different companies, it can be completely different things. And so, even though the word data science was coined, you said around 2010, people have been doing data science work for decades, maybe centuries. If you were, 100 years ago, you could have been collecting data and building a model and then studying it with a T-test and hand drawing a chart and getting that published and you’re a data scientist 100 years ago.
Kirill Eremenko: 01:07:43
Yeah, actuaries have been around for centuries. I don’t remember when, was it 19th century or 17th? Maybe 19th, the first insurance company was formed in the UK. I think it was like a church insurance company or something like that and they had to calculate the likelihood that people will die at a certain age and so on, so actuaries have been around for a long time. That’s another form of data scientists.
Jon Krohn: 01:08:16
Predictive model, for sure.
Kirill Eremenko: 01:08:19
Yep, so these are the five things that you need to get, whether you get them through our challenge, which you can find at www.superdatascience.com/challenge, or whether you put them together on your own. From my experience, I think these are components for success and maybe you’re missing one of them in your preparation for a data science career, so you can augment it with that.
Kirill Eremenko: 01:08:41
Number one is you need a plan, a weekly plan and you need to stick to it. Data science is very broad and you can easily get sidetracked into focusing on one thing or looking at one example or reading one book for too long and you’ll get sidetracked and off-track, and you will just maybe get discouraged, because so much time has passed and you’re not making progress. Number two is, you need top notch content. You need five star content. There’s lots of content out there. There’s lots of people putting content out there, which is good in general for people to share their thoughts and ideas and to help, because maybe somebody’s good at something that other people aren’t, but you need to not get lost again, not get overwhelmed by all of this content and to find one, two, maybe three reliable sources.
Kirill Eremenko: 01:09:32
“Okay, my machine learning tutorials, I’ll get here. My visualization ideas and tutorials, I’ll get here. Thoughts on how to clean data, I will get over here or the questions, I’ll look at questions people ask and the answers they’ve gotten.” So, have those sources identified, because otherwise you’re going to end up looking all over the place all the time, trying to find content and it just makes your life easier and less overwhelming if you know where you need to go, when you need to study a certain thing. Step three is, you need projects to build your portfolio. This is very important for landing the job, because you cannot demonstrate the capacity to deliver on the projects that you will be doing, if you don’t have any of that evidence to show. And if you’re building a portfolio of projects, that’s going to help you.
Kirill Eremenko: 01:10:31
So for instance, you can build a portfolio of projects on GitHub. You can build a portfolio of projects on Tableau public, you can build a portfolio for projects, even on LinkedIn, sharing your work. You can build it on Medium. You can build it on your own website. There’s lots of ways you can post your projects. And the important thing here is, try to combine this… Oh, another place is Kaggle. You can post your projects on Kaggle. You can do projects there, but try to combine your content, what we talked about in the second point with this third point, that your content provider or whoever you’re taking courses from, wherever you’re finding them, is already including bespoke projects, not your typical Titanic dataset or virginica setosa, that’s been analyzed for decades and is being seen millions of times. Interesting stuff.
Kirill Eremenko: 01:11:22
What is the NBA doing? What’s are the top movies on Rotten Tomatoes? I don’t know, what’s happening with different economic indicators for different countries. Projects like that. Something maybe you’re passionate about, maybe you collect butterflies, create your own projects where you look at butterflies, create your own project about butterflies and there’s types and get the data and try to understand that. Or this pandemic and COVID, maybe you can get some data around that and analyze that. Maybe it’s a small dataset, maybe it’s a big dataset, it doesn’t matter, but it has to be fun and interesting for yourself and also for people that are going to be reading about it, that are going to be looking at it and understanding that, “Oh, wow, this person was actually interested in this and let’s see what they found in this really cool dataset.”
Jon Krohn: 01:12:08
Yeah, I think, from the things we have been saying, that can be everything. If you can demonstrate projects, you have half a dozen or a dozen different projects and having these kinds of structured paths is helpful, because it helps you figure out, “Okay…” Unless maybe you know, you know like, “I only want to be doing computer vision and that’s the only thing that matters to me.” And then, okay, go do that, but often a really great approach is to use a structured program like the 99-day SuperDataScience challenge, so that allows you to say, “Okay, we’re going to have this vision project, language project, this visualization project.” And so, you can focus on these different aspects to be a more well-rounded data scientist. So that might help you figure out what kind of data scientist you want to be, but also might increase your job [options 01:13:00].
Kirill Eremenko: 01:12:59
Absolutely. And I’d like to be fully transparent with everybody, so with the 99-days challenge on SuperDataScience, you won’t get the content, you won’t get the courses unless you’re already signed up to SuperDataScience, or you have them on Udemy, so that is part of the SuperDataScience membership. But what you will get, is you will get this 30-day learning plan, absolutely free. You just go in, put your email in, you get the 99-day learning path.
Kirill Eremenko: 01:13:26
Some of the content that is referenced in the 99-day learning path is reading materials or other courses or other videos out there, which are available online for free, but you can also just take that learning path and that’s what we also encourage you to do, if SuperDataScience is not the place for you to learn, it’s not something that you want to do, you can just find your own content. You can find providers on YouTube of content. You can find courses, maybe you’re already taking a course. Maybe you’re already signed up to some other platform and you can compliment that learning path, that 99-day challenge, with the content that you already have access to or you prefer. So, no problem with that, the learning guide itself, absolutely all yours to take at the 99-day challenge website.
Jon Krohn: 01:14:16
That’s cool. I don’t know if we’d explicitly said that it’s free. That’s cool. I didn’t even know that. I mean, obviously then, so you’re suggesting this is a specific Udemy course you could be watching or industry produced on this platform, you could be watching this and so, that’s kind of an easy way to follow up, but that’s cool that you’re making the 99-day 30 page pack completely free, and then people can follow it in other ways. I didn’t realize that. And so, a related point in terms of a free resource that you can use to help you, if you’re looking for problems to solve datasets or example ways of solving problems that might inspire you for specific projects, I collect those on my personal website. So jonkrohn.com/resources, those kinds of things are available there for you if you’re looking for something to inspire you or a dataset to tackle.
Kirill Eremenko: 01:15:11
Awesome, that’s really cool, jonkrohn.com/resources. And also, Jon has a YouTube series on machine learning fundamentals. Check that out. That could be a good way to get your content on machine learning and Jon’s done in person talks, conference talks, webinars, podcasts, what have you not done? Book, you’ve written a book, so definitely a reliable source. Content as well. So there’s lots of places, but you need to have this whole structure. It shouldn’t be like, “Oh, I’m going to go learn data science. What I’m going to do today?” It has to be a structured plan to follow along in step by step and you need to know where to get one.
Jon Krohn: 01:15:51
Thank you. Thank you for mentioning some of my stuff there, I appreciate that. All right, number four, number four, number four.
Kirill Eremenko: 01:15:57
Number four. So, the fourth thing that is needed is a pathway to specialize. As we discussed, you can’t get all of the data science in your toolkit right away, especially. So you need to figure out what will you go into, will it be machine learning? Will it be data visualization? Will it be data wrangling? By the way, Jon, do you have any other ideas, because those are the three specializations we included in this 30 day guide, because they seem like the most obvious ones. Do you have any other ideas of what people could specialize in?
Jon Krohn: 01:16:29
Data wrangling, data visualization.
Kirill Eremenko: 01:16:31
Machine learning.
Jon Krohn: 01:16:32
Machine learning. Those are good, big ones. I mean, there’s modeling approaches other than machine learning, right?
Kirill Eremenko: 01:16:48
Like what for example?
Jon Krohn: 01:16:49
Statistical approaches to modeling data instead of machine learning, so…
Kirill Eremenko: 01:16:54
That’s a good point.
Jon Krohn: 01:16:56
So yeah, so that’s just another, so maybe kind of machine learning more broadly is kind of like modeling data and that includes machine learning, the kind of standard statistics that you learn about like regression, T-test, but then also Bayesian statistics would be part of that kind of modeling umbrella. I think those are pretty good. Those are pretty good as three main categories, visualization, modeling, data wrangling, otherwise-
Kirill Eremenko: 01:17:23
Maybe something to do with communication, like being the person that translates data insights to the business leadership?
Jon Krohn: 01:17:32
That’s a cool one for sure. That’s a really great idea. Probably related to visualization. But, I guess, there’s kind of things that maybe are less directly, this is definitely data science, but things like data engineering, so having an intimate understanding of how to build pipes in a backend to allow data flows to happen and models to be deployed. So this is kind of like, I call that a full stack data scientist if you can deploy your model into production yourself.
Jon Krohn: 01:18:06
So, there’s these kind of software developer focused, data scientist is maybe another place. And then, I don’t know, even like management. So, a data science manager is kind of a specialization where… It’d be pretty rare to start there, but it might happen. There might be some scenario, where you have a huge amount… In fact, I do know people that this has happened to. So, I know people who have been a consultant and an economist, and they’re a relatively senior person, they move into data science as like a data science manager. Maybe they wouldn’t be able to get into the weeds in a huge amount of detail with the people that report into them. But they’re really great at problem solving, and managing the corporation and getting projects done. So, I don’t know, there’s a couple more ideas.
Kirill Eremenko: 01:18:59
That is interesting, because a lot of the time I find data scientists as they grow, they’re kind of pushed into management, like nudged. Hey, maybe you should be a manager now. But, they really like the technical side of things. And then management is a whole different story, managing people, then I think you had talked about this with Erica Greene on the podcast, like you don’t have much more time left to now do the data science work yourself, so you have to kind of like split your time in half. When will I do the technical stuff? When will I do the managerial stuff? So, maybe there is room for people who are experienced managers to just learn the nuts and bolts of data science, and bring their managerial skills from a different field into data science.
Jon Krohn: 01:19:41
I think there’s… I mean, I’ve seen success in people doing that. As you were just speaking there, I had another couple interesting ideas of kind of like data science specializations. So, one of increasing importance is ethics. So, people specializing in issues around bias in algorithms, AI ethics. So, that’s something that’s kind of growing and can be separate. And so, people maybe with more of a policy or governance background could be really useful in that kind of data science specialization. And then also, I mean, kind of just more generally, so you can end up having people in data organizations, so they can be data leaders without really being data scientists at all. So, if you think about like a chief data officer, they’re responsible for maybe procuring data sets externally and managing the administrators that maintain the databases, but they don’t necessarily ever do any modeling or visualization themselves. So I don’t know, just a few other specializations that come to mind.
Kirill Eremenko: 01:20:53
If anybody listening to this manages to take the 99 day challenge and become a chief data officer, please email us, let us know. That would be really cool, like just starting into the field and straight to chief data officer.
Jon Krohn: 01:21:07
I mean, there [crosstalk 01:21:07] must be more junior data officers too. I guess, that’s like a database administrator kind of role.
Kirill Eremenko: 01:21:18
Data custodian.
Jon Krohn: 01:21:19
Data custodian. Yeah.
Kirill Eremenko: 01:21:22
I think we had a role like that one [inaudible 01:21:24]-
Jon Krohn: 01:21:24
Oh, Yeah. Really.
Kirill Eremenko: 01:21:25
Yeah. At the separation company. Okay. So-
Jon Krohn: 01:21:30
One last one, right?
Kirill Eremenko: 01:21:31
Number five. Last one. Very important. Have a career preparation checklist. A lot of people go through, all right, I’m going to learn data science, I spent 99 days or 150 or six months, whatever. Then they’re like, “Okay, I’m ready.” “I got this under the hood.” I don’t even know if that’s the saying, under the hood. “I’ve got this under my belt.” [crosstalk 01:21:57] “I’ve got all the data skills, I’ve got the projects, my portfolio,” and so on. “Now, I’m going to start looking for a job.” And then, the whole process starts from there, looking for a job. Why not combine? That’s the best way to do it. Spent two and a half hours learning data science a day and 30 minutes doing something for your career.
Kirill Eremenko: 01:22:18
There’s things like reaching out to recruiters, reaching out to people in the industry, networking with people, attending events, attending free conference, attending paid conferences. Updating your CV and especially LinkedIn. Putting content out there. Commenting on other people’s posts. Looking at jobs and comparing job descriptions to the things that you’re learning. Looking at different types of job descriptions. Looking… Comparing them against each other. A lot of it is about networking and getting yourself out there, because I think there’s a statistic, it’s quite old, I don’t know if it’s still the case, but I’m pretty sure it is, that 70% of jobs get filled before it been advertised. Through connections, through networking, and only 30% are like direct, okay, apply here, here’s the resume and so on. So, it also has its room, but the more of these things you do along the way, you have 99 day, so you can get a lot of this stuff done. You can get a lot of ground work, cover a lot of ground, so by the time you get to day 85 or day 80, you already have all these connections.
Kirill Eremenko: 01:23:32
You already have some people interested in your work, some people following you, you’re answering some questions, some people answering your questions. And then, it’s much easier to go and like land a job, or see what opportunities exist. And moreover, new opportunities will open up. If you just start looking for jobs at the 99 day mark, like from day 100 onwards, or whenever you’re ready, you’re going to find jobs that are available then. But if you’re looking for jobs along the way, you’ll have more exposure, and you would have seen more jobs that pop up and disappear, and you will know what to expect and what would you personally are interested in the most.
Jon Krohn: 01:24:09
Got a question for you Kirill. What’s the difference between number one, the plan, and number five, the checklist?
Kirill Eremenko: 01:24:17
Number one, the plan it relates to both the content part, like what you want to learn and also the career. So, you need to have a plan for both. The career preparation checklist, it’s kind of like to reiterate the specific steps that you need to follow and include in your plan.
Jon Krohn: 01:24:39
Yeah. It must just be… It’s just, we’re distinguishing clearly the plan and the checklist. The checklist is making sure that you’ve executed on the steps on your plan, or like…
Kirill Eremenko: 01:24:53
Well, let’s call it this way. The plan is, which content am I going to learn? Okay, I want to do intro to data science, basics of visualization, coding for data analytics, whatever, data visualization. You need to write out predominantly, what are you going to study and in which order? Whereas the Career Prep Checklist has nothing to do with content, it’s all about your career steps.
Jon Krohn: 01:25:19
Ah, nice. I got you.
Kirill Eremenko: 01:25:23
Yeah.
Jon Krohn: 01:25:23
Thank you.
Kirill Eremenko: 01:25:25
No worries.
Jon Krohn: 01:25:26
That sounds brilliant. I was convinced before we even got into all of this level of detail. As I said at the onset of this episode, as soon as you mentioned to me that it could be possible to get a data science career in 99 days, these kinds of steps that you’ve outlined over the course of this episode are the kinds of things that immediately came to mind for me, as a way that you could do it. And, I think it’s brilliant that you’ve gone through it in all this detail. I think it is awesome that you’ve set it up in a way that people can just go to super datascience.com/challenge and download the guide for free. It’s really brilliant. I don’t know if you have anything else you want to add before I ask you for a book recommendation.
Kirill Eremenko: 01:26:11
I would like to add that there’s a saying, “Success is 80% psychology, 20% mechanics.” So, how you do something is less important than the reason behind why you’re doing it. So, before jumping into all of this, make sure you have your why very well identified, that there’s a reason why you’re doing this and it’s a strong reason that speaks to your heart. Maybe you want to change the world, maybe you want to change your life circumstances. Maybe this is something you’re passionate about. So, have the reasons why identified, set yourself up for success by outlining these steps that you’re going to take. Take a day, take two days, take a week if you need too, just to prepare for this journey, for this challenge, or however are you going to do it. Don’t just rush into it, and get lost and get overwhelmed.
Kirill Eremenko: 01:27:19
But set yourself up for success by doing some prep work first, and then come the mechanics. And then, just execute. Execute for 99 days, or if you have less time, for 150 or 300 days, but execute. But if you get those two things right at the start, if you get the psychology right and the vision, the strategy, the preparation in the first few days, then it’ll be much easier to execute and you’ll have a target that you want to hit. And, it’s just a matter of putting in the work together.
Jon Krohn: 01:27:55
Beautiful message and so important. You need to have the why for sure. To be sure. Hopefully, lots of people who’ve been listening to the podcast have developed their why’s. But good to check in on yourself and make sure. All right, Kirill, you know the last question. What’s your book recommendation?
Kirill Eremenko: 01:28:11
Book recommendation. Amazing book, I’m listening to audible right now. It’s…
Jon Krohn: 01:28:17
He’s listening to it right now, as he’s doing the podcast.
Kirill Eremenko: 01:28:22
Maybe. Maybe when you listening to this, I’m listening to the book.
Jon Krohn: 01:28:25
He’s done so many podcast episodes.
Kirill Eremenko: 01:28:27
Yeah.
Jon Krohn: 01:28:28
So, Kirill had done 431 podcast episodes before I took over at the beginning of this year. And so, you can just go on automatic pilot. This whole time that he’s been doing the podcast, he’s been listening to a book.
Kirill Eremenko: 01:28:43
I wish. You’re getting there as well. What are you on, like episode 40 almost or around 40?
Jon Krohn: 01:28:50
Well, I have to do some subtraction, but I think [crosstalk 01:28:54] we’ve recorded more than 40 now.
Kirill Eremenko: 01:28:58
Good stuff.
Jon Krohn: 01:28:59
Yeah.
Kirill Eremenko: 01:28:59
Good stuff. Congrats, man. How do you feel?
Jon Krohn: 01:29:03
It definitely gets easier. I mean, I was never stressed out about it, but I always… You did a really great job of coaching me. Kirill created for me, there’s no way the audience would know this, but I don’t know off the top of my head how long the document is. This handover document that’s like 20 pages. And, I remember when he sent it to me, he spent like a weekend making this document. And then, it took me like a day to read it and go over everything, and understand that there’s a huge amount of process that goes on behind-the-scenes to create these well-oiled episodes that you experience. There’s so many steps that Kirill had mastered over the years and he conveyed those to me. And, I learned from him for a few episodes. I learned from you, Kirill and you made it easy. There’s never any point… Maybe actually, the only time I was nervous was the very first one, when you and I co-hosted an episode.
Jon Krohn: 01:30:01
So, that one aired at the end of December, he was Syafri Bahar. And, that was the only time that I showed up and I had kind of like butterflies. But, it’s been smooth sailing and it gets easier. The production and editing team, the management team that you put together makes it so easy. They’re utmost professionals, unbelievable. And, they just… Being the host of the show is a piece of cake.
Kirill Eremenko: 01:30:27
Thank you. Thank you for the kind words, Jon. And, also on behalf of our listeners, students, I want to thank you too. You’re changing lives. You’re helping thousands of people every week. You’re putting in your personal time into this and that’s worth a lot. So, thanks a lot.
Jon Krohn: 01:30:47
It is a huge joy to be able to do this. It is an honor. So, thank you.
Kirill Eremenko: 01:30:53
Awesome.
Jon Krohn: 01:30:54
And, thank you everyone in the audience for allowing me to have… Even if you weren’t there, I’d be sitting here and doing this episode with Kirill, and we’d just be sending it into the ether and no one would listen. It’s so much fun to do. So, to have people actually be out there listening and to reach out, to get these messages that people share on social media. And, it’s absolutely wonderful. So, thank you so much Kirill for the opportunity. All right. [inaudible 01:31:21] let’s get to your book recommendation.
Kirill Eremenko: 01:31:23
Book recommendation-[crosstalk 01:31:23]…
Jon Krohn: 01:31:23
I’ll stop you there.
Kirill Eremenko: 01:31:26
I’m going to show it on the video version. It’s called The Courage to Be Disliked. Can you see there?
Jon Krohn: 01:31:33
Oh yeah.
Kirill Eremenko: 01:31:34
Okay. So, it’s a book that was originally written in Japanese. It’s called the… Or the subtitles, The Japanese Phenomenon That Shows You How to Free Yourself, Change Your Life and Achieve Real Happiness. It’s been translated into English, not into many languages. I wanted to get a copy in Russian to give to my mom, but unfortunately its not available. Maybe somebody will translate it. Its sold 3 million copies.
Kirill Eremenko: 01:31:59
The authors are Ichiro Kishimi and Fumitake Koga. Amazing book, I’m about halfway through it. I think it’s about eight hours long, if you listen to it on audible. And, it talks about Adlerian psychology. So, the psychology we are familiar with in the world is mostly by Freud and his student, Jung. Well, Freud had a contemporary called Adler. And Freud in psychology is, you look at the past. Somebody is sitting at home, has social phobia, can’t go out. Well, they probably had a trauma in the childhood. Maybe their parents were abusive, or very controlling, or very manipulative or something like that. So Freudian and Jung psychology tend to look at the past, to look at the trauma that happened, so that you can get in and help a person resolve that trauma to be able to move forward.
Kirill Eremenko: 01:32:54
Adler on the other hand is less well known. But I find his principles, the ones described in a book, very interesting. Because, they look at the future. They say that the reason for anything happening now is an objective that you have set yourself in the future. And, the example of somebody sitting at home and having social anxiety, or not wanting to go out, as controversial as it might sound, in Adlerian psychology would be explained that, that person is doing that because they have some objective in the future. And, what objective would they have? Well, let’s look at how will this person’s parents feel if this person is sitting at home and is not able to go and socialize? That person’s parents are going to be worried and they’re going to show him attention, or her attention. And, they’re going to want them to be better and they’re going to check up on them.
Kirill Eremenko: 01:33:46
And so, they will be getting attention, love in a certain way and care from their parents. If they were to go outside, then they will become average or even less than average out there. And, their parents would be happy that they’re out finally. And so, they wouldn’t be fulfilling this objective of attention from their parents. It’s all, of course, not conscious, it’s subconscious. But, the point is that Adlerian psychology, in my perspective, allows you to address some of your issues as more of a choice, rather than a consequence of the past. That you can, if you really think about it, in my view, I’m not a psychologist and I don’t want to undermine anybody’s problems, I know people are going through difficult things at different times. But according to Adlerian psychology, what my take away is that anything, any problem that I have, for instance, I’m more introverted or that I don’t like big crowds, or that I don’t know, like I have some fears, I can address them by thinking about what objective is that helping me fulfill in the future.
Kirill Eremenko: 01:34:53
So, I just need to ask myself, is that objective valid? And, how can I make a conscious choice now to change my life in the future, so that I can rid myself of these fears, anxieties and worries, or control controlling aspects of my personality or whatever else? And, I think it gives me more hope and control that I can change, that I can be better. And, it makes a lot of sense. The book is structured in a interesting way. It’s a conversation between a youth, like a young teenager, or I don’t know, a young person in their twenties with a philosopher who follows Adlerian psychology. And, is constantly challenging the philosopher and the philosopher is always constantly explaining all these questions that I would have had if I were reading this book, or studying Adlerian psychology myself. Long rant about it, but it’s a book that’s been one of the top books I’ve read in the past years. I highly recommend to anybody who’s interested in psychology and bettering themselves.
Jon Krohn: 01:35:52
Well, it sounds great. I am really interested to read that because something that I… I mean, as you know, and listeners who maybe listened especially to episodes that aired in January, February, I guess, I talked a lot about meditation. And the thing that you’re describing here, I often have this sensation, this realization that so much of what I do is driven by a desire to be liked, to do the thing that I… Like to meet others expectations. And, I don’t know if that’s exactly what this book is about, but [inaudible 01:36:32]-
Kirill Eremenko: 01:36:32
It’s a big part. The book says, “Being recognized by others should be zero of your concern, zero.” That is the responsibility of others. That is a task, they call [inaudible 01:36:45] task of others. Your task is like not wanting to be disliked. Everybody wants to be liked, but whether other people like you or not, not your problem. And the way we raise children in our society is like, okay, so if you do your homework, you’re going to get time to play computers, or are you going to get some treat. If you misbehave, you’re going to be punished or you are not going to be liked, I’m not going to like you as a parent. That is according to Adler, that was one of the biggest, like a huge flaw in society that we bring up young little people, we grow them into big people, with this whole notion that you need to be doing right things to be liked by your parents. Otherwise, you’re going to be disliked.
Kirill Eremenko: 01:37:29
So, we bring that into our adulthood and that’s what we portray. Like I find myself very often, people pleasing. I want everybody to be happy. I want everybody to like me. And, often it comes at the detriment of my own wellbeing or my own needs and desires. And Adler was hugely against that, hugely opposed to bringing kids up like that. And, this book gives you some of the answers on how you can combat in your self issue, seeing that trait.
Jon Krohn: 01:37:58
Sounds huge. All right. What a great way to end the episode. I’m sure that tons of content in this episode will resonate with listeners. And, that’s just one more piece. [crosstalk 01:38:11].
Kirill Eremenko: 01:38:10
Can I just repeat the name of the book, just so people…
Jon Krohn: 01:38:13
Absolutely.
Kirill Eremenko: 01:38:14
… maybe missed it. It’s called The Courage to Be Disliked.
Jon Krohn: 01:38:18
The Courage to Be Disliked. Great. Thank you so much Kirill for coming back on the SuperDataScience podcast. I can’t wait to have you on again soon. Obviously, you and I will be in touch, and I’m sure it’ll be no time before you’re back here again and sharing your pearls of wisdom with us.
Kirill Eremenko: 01:38:35
Thank you very much, Jon. Appreciate you having me, and even more so I appreciate you hosting all these episodes. I’ve listened to a couple of them. It’s always a lot of fun to hear how you’re interviewing this. Thanks a lot.
Jon Krohn: 01:38:46
You’re most welcome. All right. Catch you later.
Kirill Eremenko: 01:38:50
See you.
Jon Krohn: 01:38:56
Fun catching up with Kirill in this episode, he’s such a wise person. I personally always gained so much from my conversations with him. I hope you too learned a lot from Kirill, particularly about getting started in data science, as well as the psychology of career choices, such as the five myths that erroneously lead people to think that they cannot obtain a data science job. At the end, we wrap the episode up by getting filled in on the five things you need to land your first data science job. That’s a plan, top-notch content, projects for your own portfolio, a specialization and a checklist. To get you rolling on all five of these items straight away, you can check out Kirill’s 99 day path to landing your first data science job at super datascience.com/challenge. As always, you can get all the show notes, including the transcript for this episode, the video recording, any materials mentioned on the show and the URLs for Kirill’s LinkedIn profile, as well as my own LinkedIn and Twitter details at www.superdatascience.com/471. That’s super datascience.com/471.
Jon Krohn: 01:40:08
I’m always delighted to meet listeners, so please do connect and feel free to tag me in posts with your thoughts on the episode. Your feedback is invaluable for figuring out what topics we should cover on this show next. All right. Thanks to Ivana, Jaime, Mario and JP on the SuperDataScience team for managing and producing a very special episode today. Keep on rocking it out there folks and I’m looking forward to enjoying another round of the SuperDataScience podcast with you very soon.
