Jon: 00:00:00
This is episode number 623 with Shashank Kalanithi, Senior Data Engineer at BetFanatics. Today’s episode is brought to you by Datalore, the collaborative data science platform.
00:00:14
Welcome to the Super Data Science Podcast, the most listened-to podcast in the data science industry. Each week we bring you inspiring people and ideas to help you build a successful career in data science. I’m your host, Jon Krohn. Thanks for joining me today. And now let’s make the complex simple.
00:00:45
Welcome back to the Super Data Science Podcast. We’ve got the sensational Shashank Kalanithi on the show today. Shashank has an exceptional YouTube channel with over a hundred thousand subscribers and it’s focused on helping people break into a data analyst career. He also has a day job as Senior Data Engineer at Digital Sports Platform Fanatics. He was previously data analyst at the luxury retailer, Nordstrom and several other firms. He holds a degree in chemistry from Emory University in Atlanta. Today’s episode will appeal primarily to folks who are interested in becoming a data analyst or interested in transitioning from a data analyst role into a data science or data engineering role. In this episode, Shashank details how you can land an entry level data analyst role in just a few weeks, regardless of your educational or professional background.
00:01:31
He talks about the hard and soft skills you need to progress from a junior data analyst to a senior data analyst position. What it takes to transition from data analyst to a typically more lucrative role as a data scientist or data engineer. His favorite resources for learning, the essential skills for data scientists and what he looks for when he is interviewing candidates. All right, you ready for this jolly episode? Let’s go.
00:02:01
Shashank, welcome to the Super Data Science Podcast. I’ve been watching your YouTube videos a lot lately and so now I’m excited to be able to interact with you in this virtual recording. Where in the world are you calling in from?
Shashank: 00:02:17
Thanks for having me here, Jon. I’m calling in from Seattle, Washington right now.
Jon: 00:02:21
Oh yeah, nice. And in the video recording, we’ve got a beautiful view out your window as the sun comes up. When we first joined on this call, it was kind of dark out there, so thank you for very early in the morning agreeing to do this call to squeeze it in before your workday. I understand that you’re not much of a morning person.
Shashank: 00:02:42
No, no, I am not, unfortunately. I find it very easy to stay up until 4:00 AM but very difficult to wake up before nine, quite honestly.
Jon: 00:02:52
There you go. Well, the sacrifices that you’re making for the Super Data Science audience, I’m sure they appreciate it out there. So, we know each other through Ken Jee. who was in episode number 555 of the Super Data Science podcast. Ken, obviously an iconic character in the data science community. He leads a great content creator agency and so that helps get big ad campaigns distributed across a broad range of different social media channels at once, which is awesome. That company’s called Learn Media, they’re doing great work. And Ken also has an enormous YouTube channel and maybe that’s how YouTube connected initially because of your enormous YouTube channel. How did you guys meet initially?
Shashank: 00:03:41
I think he reached out to me I believe, and that’s how we met. I think he reached out to me for Ken’s Nearest Neighbors.
Jon: 00:03:49
The Ken’s Nearest Neighbors podcast, which I can definitely recommend. He’s had a lot of great guests on the show. And yeah, I’ve actually reached out to him a couple of times now to say who are some of the best guests you’ve ever had on Ken’s Nearest Neighbors and Shashank, that’s how you ended up on the podcast today.
Shashank: 00:04:07
Oh, that’s an honor to hear that. Actually, I’ll be on this podcast again. You’ll probably see it around the time of this recording as well.
Jon: 00:04:14
Oh, nice. Looking forward to checking that out as well. So, your YouTube channel has over a hundred thousand subscribers, which is crazy. And on that channel you teach primarily about data analytics to aspiring and beginning data analysts. What motivated you initially to start your channel and teach about data analytics?
Shashank: 00:04:37
That’s a good question honestly. So, it really started a long time ago. So, I used to work at Interstate Batteries back in 2018 to 2020, I believe. Their North America’s largest distributor of [inaudible 00:04:51] batteries, basically car batteries. And I was working on an analytics team and supply chain over there. And then we noticed that it was difficult to get Tableau adoption in the company because people just didn’t know how Tableau worked. And so luckily my director and my boss were both very supportive of me and they helped me, I created a Tableau course, taught it to the company and just kind of had that material ready. Then March 2020 comes along, pandemic starts, and then the company my mom worked at the time was their whole business model was basically just invalidated. They did events and stuff like that. So, I mean have zero revenue for two years basically. And so unfortunately she got furloughed, no fault of her own, just wrong place at the wrong time.
Jon: 00:05:37
What are you going to do? Yeah.
Shashank: 00:05:38
Yeah, exactly. No one blames the company. What else were they supposed to do? So, I was like, okay, well what I learned was that having Tableau skills, just having Tableau skills was enough to drive a pretty solid amount of value at the company I was at the time. And so I took the basic materials I had, I turned them into a video, put them on YouTube and taught my mom. I was like, hey, here’s how Tableau works now you can put it on your resume.
00:06:05
I tell most people, you can learn Tableau in a weekend, quite honestly, it’s meant to be easy to use. So, also approach it for anyone that’s working in BI or BA over there, because I used to work in BI. If you are fighting with Tableau to get something done and it’s taking a long time to get something done, that probably is not the best way to do it. It’s supposed to be an easy to use tool, but they want you to use it a very specific way. So, I would do everything to use it that way because it’s not worth fighting with Tableau.
Jon: 00:06:32
BI is business intelligence, right?
Shashank: 00:06:34
Yes, sir.
Jon: 00:06:35
Got it. So, all right, so you were inspired to create your channel by this situation that happened with your mom. Like this opportunity to teach her Tableau, so you created some videos, you created video course or something.
Shashank: 00:06:49
Yeah, so that was like part one of the story and then for a year I didn’t really post anything after that. Then I was watching Ali Abdaal one day and he’s like, if you want to get on YouTube, just get on YouTube. He had this video where he literally said, just get on YouTube. He’s just put something out there. And I was watching this at 10:00 PM at night and I’m like, okay, well I’m not literally going to do something right now. I’ll wake up in the morning and I’ll do something. So, I woke up the next morning, I went ahead and I just did this video where I parsed out SurveyMonkey data in Python that I had to do for a client because I did independent consulting at the time. I was one of those two jobs people during the pandemic. Not a bad gig was, I liked it, I don’t regret it at all quite honestly. But when I had just showed myself actually doing that and I titled the video Day in the life of a Data Analyst.
Jon: 00:07:35
That was a good video. That video has 2 million views, 2.1 million views at the time of recording. That’s wild. That was one of your first uploads?
Shashank: 00:07:47
Yes, yes.
Jon: 00:07:50
That’s incredible.
Shashank: 00:07:54
There was a little bit of thought put into it. The basic thought was, so back in 2015 – 2017 Mayuko Joma Tech, it’s like a really big YouTubers, they kind of said day in the life of a data scientist day in the life of a software engineer and back then seeing the day in the life of a FAANG company. So, Facebook, Apple, Amazon, Netflix, Google. I refused to call Facebook Meta, it’s a dumb name. But they used to work at those companies and they would show a day in their life and it was pretty cool.
00:08:27
But then what happened was people started to realize that you could get just a ton of views by just doing that. Being like, hey, I don’t actually do any work. Here’s my coffee machine, here’s my all the cool stuff that Google gives us. And then not show any actual work. And so this video is a bit of a shot at the newer versions of those videos. Again, if you’re the first to do something, it’s pretty cool, but if you’re number 10,000 in a line of a hundred thousand people waiting to do something, then I don’t really see the value being added. Everyone knows that FAANG employees are some of the best treated employees in America right now. We want to see the work that you do. So, that was how that video came about.
Jon: 00:09:04
Cool. Well certainly doing well.
Shashank: 00:09:06
2 million views is all luck. Yeah.
Jon: 00:09:08
Yeah. I mean I’ve released a lot of videos and I don’t think all of my videos together are at a fraction of 2 million. I mean, to some extent it’s luck, but it also is, you thought about what kinds of videos are getting traction and you created this video as a specific statement, an antidote to those kinds of FAANG videos. So, there was a good insight behind there. It’s not just luck and obviously there had to be some high quality in there in order to have that get such a big audience. So, you must have, or you do have, I could say an innate ability to identify high-quality video content. You’ve got a good sense for the art. All right, so back to your story about how you got started before that 2 million view video where you were talking about creating that Tableau course. Do you think that anybody who’s out of work could become an analyst? Should anyone who’s out of work become an analyst?
Shashank: 00:10:12
So, as far as like can anyone do it? I think anyone can do it. If you want to do it, you can definitely do it. And this isn’t just me saying that. If you want to be in the NBA, you can’t do it. Like desire is not enough. If you want to be a PhD level scientist, 99% of you do not have the capabilities to do so, myself included. So, I’m not just BSing you here and saying, oh yeah, if you have the willpower, you can do it. No. No. This is actually something that you can do. Because the level of technical skills necessary to break into the industry are relatively low actually. I teach a bootcamp where we do hammer in Python and SQL and Tableau and stuff like that. But that’s kind of to over prepare people. But as far as if anyone wants to do it, they totally can do it.
00:10:58
Now should anyone do it, I kind of feel like basic Tableau skills or BI skills, so Power BI, same thing. Again, given that you can learn it in a weekend, I believe it benefits almost everyone who works a knowledge job. What do they call it? Knowledge economy job. Because you would be surprised how many decisions are made in corporate America. I mean, just giving corporate America as my experience that don’t involve any or much rigorous analysis of data. And if you’re the first person to bring in that rigorous analysis of data through just a BI tool basically, which allows you to measure your metrics on a regular cadence and allows you to transform datas in interesting ways and present it to people so that they may understand the insights that you have. The ability to do that is something that’s going to accelerate your career in the industry.
00:11:54
So, I’m a believer that BI skills, I think everyone should know. Basic SQL is helpful to a lot of people. The number of business people I know who know SQL are very, very low. But the ones that do are always able to get access to databases because engineers and analysts are too busy to run every single query that could possibly be run. So, eventually everyone just gets frustrated and they’re like, oh, here’s database access. Go run your own queries. And again, then as a business person, you’re able to run the kinds of analysis that your coworkers can’t do and you make yourself more useful.
Jon: 00:12:29
Yeah, in just a weekend.
Shashank: 00:12:30
Exactly. Exactly. You can learn, maybe not SQL, but you can learn Tableau on a weekend easy.
Jon: 00:12:34
Yeah, that is an awesome tip. And so your YouTube channel actually is filled with tips. So, even though it’s only been a year since you started creating a lot of content, there are a lot of videos there, you’ve got tips for data analysts obviously as we’ve already discussed, so things like resume advice, interview prep advice, you have chapter walkthroughs of iconic data analytics, data science textbooks. You have guidance on software. So, how to use the Notion API with Discord and Python, laptop hardware decisions. So, you’ve got quite a variety there. What inspires you to make your particular topic choices when you’re thinking about creating a video?
Shashank: 00:13:19
Usually it is something I see missing in my YouTube channel. It’s something that I just happen to pick up and then I want to teach people. I usually pick something up, learn it myself, and then try and teach people how to do the same thing in a very practical way. And oftentimes it’s inspired by questions. So, one of them was the laptop hardware video where my opinion on laptop hardware is that it doesn’t really matter. For 90% of analytics tasks, any basic laptop will do. Just make sure you have at least eight gigabytes of RAM and a descent CPU, which you’d have a hard time buying a bad one in today’s market.
00:14:00
So, a lot of my videos are an effort to try and tell people that stuff is easier than it seems and here’s the fastest way to get into it. Here’s the fastest way to get a job with these skills. I try and always coach or couch whatever I do in the context of getting a job because I wouldn’t call myself an academic by any means. I like learning, but I’m not an academic. I like doing activity and doing work that drives some specific economic value, aka going to the office and stuff like that. And so I really couch my videos in those terms. What can you do? What skills can you learn to quickly build up your ability to get a job. And then when you’re in a job, that’s the best practice you can possibly get. That’s usually how I think about what videos I come out with.
Jon: 00:14:53
Today’s show is brought to you by Datalore, the collaborative data science platform by JetBrains. Datalore brings together three big pieces of functionality. First, it offers data science with a first class Jupiter Notebook coding experience in all the key data science languages, Python, SQL, R and Scala. Second, Datalore provides modern business intelligence with interactive data apps and easy ways to share your BI insights with stakeholders. And third, Datalore facilitates team productivity with live collaboration on notebooks and powerful no code automations. To boot with Datalore, you can do all this online in your private Cloud or even on Prem. Register at datalore.online/sds and use the code SuperDS for a free month of Datalore pro and the code SuperDS5 for a 5% discount on the Datalore enterprise plan. So, get the job and then make a commercial impact early once you get the job.
Shashank: 00:15:49
Exactly, exactly. Because I think a lot of people, you remember that show, Are You Smarter than a 5th Grader?
Jon: 00:15:58
Yeah.
Shashank: 00:15:58
So, that show came out when I was in fifth grade and it was really funny because I remember, oh yeah, how could an adult not be as smart as a fifth grader. Now being an adult, I totally get it. You just forget 90% of the things you’ve learned in school. Not to devalue the importance of school, but to say I’m really focused on teaching people stuff that they can keep for the rest of their lives. Once you learn how to program once, hopefully you never forget it again. SQL, Python, Tableau, the basic framework that you use to analyze what laptop should I pick, this kind of stuff is skills that are useful for the next 20 to 30 years of your life maybe.
Jon: 00:16:36
Great. Yeah. So, it sounds like you’ve identified a way of thinking of videos that are going to be practical and are going to provide people with tips that they can use for years and years to come. Sounds hugely useful. So, across your videos, all the guidance that you have on the skills that are needed to be a data analyst, what do you think are the top technical skills that are must-haves to be a data analyst?
Shashank: 00:17:06
That’s an interesting question. I think, and when I say must-haves, let me also be clear that there are jobs that don’t require these skills and there’s some jobs that you can go into that they’ll teach you the skills from day one. Here’s a better way to put it, the must-have skills to be a senior data analyst. Basically if you want to advance in the industry, what do you have to have? I would say some BI tool, Tableau, Power BI, they’re basically all the same. SQL, you have to know SQL and then Python is a very nice to have where you do… No data analyst job I’ve seen required that you know Python. But you can convince people to pay you more and you can drive more value. If you can drive more value that’s how you convince people to pay you more by knowing a programming language like Python.
00:17:58
And then Excel, I would say PivotTable, VLOOKUP, HLOOKUP. Again, this is all stuff that can be learned in the weekend. So, out of all of those skills, SQL’s the only one that you can’t learn in one weekend. So, that’s why I think anyone can be a data analyst because Tableau and Excel are things that you could give someone a weekend and just tell them, don’t go out with your friends, don’t go out, don’t do anything else. Just study this for a weekend.
Jon: 00:18:22
So, one weekend of Tableau, another weekend of SQL, another weekend of Excel, and then you’re in pretty good shape for your data analyst interviews. And then once you get that data analyst job, you might want to start working on your Python and developing that out either on the job or maybe on your weekends once you start.
Shashank: 00:18:42
You might need two weeks of SQL, but yeah.
Jon: 00:18:46
Nice. Two weekends?
Shashank: 00:18:48
I’d say two weeks.
Jon: 00:18:48
Two weeks, okay.
Shashank: 00:18:50
Yeah. What I’ve learned is that there’s this really big mental barrier and I used to have it too. There’s a mental barrier for people when they see text on a screen doing things instead of just saying things, right? 90% of the time text on a screen is just to say something to you, warning message here, Microsoft Office document over here. But when people start putting in commands into a computer to start telling the computer, when you start talking back to the computer that way there’s some mental barrier. It’s a weird mental barrier that people have. And I 100% had that I think stops people from learning how to code. Because the basic of coding are very, very simple. They’re not hard at all. You could teach a fifth grader how to code in Python over a weekend at the basic level.
00:19:31
Now environment setup and stuff like that. Yeah, no, that takes a PhD level person to figure out environment setup’s absolutely miserable. But the reason I say SQL takes two weeks is because that mental barrier takes a while to get over. Some people get over it instantly. Some people who, if you’re exposed to technology your entire life, you may be able to get over it more easily. But I keep kicking myself for not learning how to code in Python earlier. There’s so much stuff I could have done when I was a teenager had I known how to code in Python. But eventually you get over that barrier and then you’re able to go from there. That’s why I would say about two weeks for SQL.
Jon: 00:20:05
Nice. And SQL is a pretty good gateway drug for getting people into programming because it does so structured querying language. It does have a bit of a human language feel to it because the way that you make commands, you’re select these columns from this database. It’s quite a natural feeling thing. You can convert the SQL language typically quite easily into human natural language. All right. To summarize those, the must haves for being a senior data analyst are a BI tool like Tableau, SQL, and Excel and then Python is hugely valuable, is a nice to have as you get into that role. All right, so that’s hard skills. What about soft skills? What soft skills are needed to Shashank in the analytics space?
Shashank: 00:20:55
Honestly, Jon, I’m glad you asked that. It’s a question that’s not asked enough. And I think it’s because soft skills … One, soft skills, as long as you’re not a serial killer, no one cares about your soft skills for an entry level job in the technical fields. So, I went to Emory, which is a heavy business school. A lot of my friends are B school people. For them, soft skills are very important. And also the funny thing is it also created a very high bar for what I expect out of people from a soft skills interviewing perspective, which unfortunately the engineering world has disappointed me with.
00:21:29
But as far as what soft skills do you need? So, you can get into a company basically purely on your technical skills at an entry level. If you want to advance your career, then you’re going to need the soft skills. So, one of the big ones is requirements gathering, which is the skill of going into, or hearing a problem from someone and being able to deep dive into what are the actual requirements we need, we have to have in order to solve this problem. That’s a really big one. One that is needed for basically any senior position in any company is project management. Because you as a data analyst are going to have to have engineers and data scientists do stuff for you as well. And making sure that they actually get that stuff done in a timely manner is a very important skill to have.
00:22:16
So, project management is another really big one. Presentation skills are a big one as well. Prioritization. And prioritization is a very interesting one because you’re going to have to know how to prioritize tasks depending on who gave them to you. I’ve been in more than one situation where I’ve been given multiple tasks and only one of them could get done in a day. But one was given by director and one was given by someone who wasn’t a director. And even though it have made more business sense to do the task an non director gave me, it gave a lot more visibility in the company if I did what the director asked me to do. And that’s kind of playing politics, that’s up to you how much you want to do that personally. If you want to advance past a certain level, past a senior level, you do have to play politics.
00:23:06
That’s why they call senior analysts, senior engineers, these are what you call terminal positions, i.e they’re the position that someone could stay at for 10 years and no one would blink an eye. They’d be like, oh yeah, that guy … He’s been an analyst for 10 years, a senior analyst for 10 years. If you were an entry level analyst for 10 years, you would’ve been fired eight years ago. But because there’s certain positions you have to advance, otherwise people are like, what is this person doing? And so if you want to advance past that terminal position, then these soft skills become extremely important.
Jon: 00:23:38
All right. So, you mentioned that the soft skill bar that you have to clear for an entry level data analyst role is pretty low. You’re saying that it’s basically nonexistent, but then once you do get that entry level data analyst role to become more senior, there are skills like requirements gathering, project management, presentation skills and prioritization that are all critical to becoming a senior data analyst. And then you described a senior data analyst position is potentially being terminal. You could be stuck there for a decade. But you Shashank you transitioned from being a data analyst two years ago to a role at Nordstrom where you got to do a lot of data science work.
00:24:17
And then most recently you became a data engineer for a digital sports company called Fanatics. So, what is the difference between these roles, data analyst, data scientist, data engineer? And it seems like there is a reasonably common progression from analyst, scientist to engineer. So, how did you make that transition? Why is that transition so common and what tips do you have for listeners to get out of a terminal senior data analyst position into one of those other kinds of roles?
Shashank: 00:24:54
That’s a very interesting question. One more soft skill I would add. If you’re entry-level, willingness to learn, you need to be able to show that very willingly. I feel like I need to add that in there because the thing is when you leave college, you’ve spent the last, what? 16 years. I’m assuming bachelor’s degree. You spent the last 16 years being schooled and you would hope you come out of that knowing a bunch of stuff and being able to show your knowledge. That’s not what people are looking for from entry-level candidates. They’re looking that you are willing to learn because it turns out those last 16 years aren’t actually that useful. Your learning starts today. So, one more soft skill I’d add there. Kind of going to the question you just asked about transitioning from data analyst to scientist to engineer and why that would be a common path.
00:25:38
I think for me, I don’t know if it’s a super common path. I know of only a couple of people who have made that transition who don’t have master’s degrees. And people who go from science to engineering. That transition I’ve seen, I’ll kind of get into that in a second. So, we’ll go from analyst to scientists first and talk about that. So, most people who I know who are data scientists have master’s degrees. It is a highly, I would say a mentally rigorous position. And it’s a position, like being a data scientist is a role where you’ll spend maybe 60% of your day just deep in thought. You need to be able to think deeply for long stretches of the day. More so than engineering, more so than analytics. That’s what makes a data scientist, a data scientist. And I’m kind of guessing that’s why the master’s degree is a soft requirement to get into the field.
00:26:31
Now I don’t have a master’s degree and I didn’t really intend on getting one after, at least for data science. After I talked to a bunch of data scientists and they were like, I mean yeah it was helpful to get past a recruiter. But I don’t use my master’s degree at all in this job. And then I looked at the work they were doing and I was like, no, I mean I can do that work. I don’t need a master’s degree for that. So, I’m not going to go. Even if the company pays for my degree, it’s a lot of my time and effort that’s spent doing that, that I could spend doing something else.
00:26:58
And so the transition was helpful because data analysts do not require a master’s degree at all. And it allowed me to get into the company and it put me in the same basic zone as a data scientist that allowed me to pick up work by myself that was able to show my boss, hey look, I’m doing data scientist work. Because you have access to all the databases, you can get computers that are unlocked. That’s a really big one. A lot of business people, they can’t install apps on their laptops. Engineering orgs allow people to install apps on their laptops, otherwise they just wouldn’t be able to get work done. So, being an analyst, you get that kind of a laptop. So, I was able to just install Python on my laptop and show them a bunch of stuff I was doing, which is probably why that transition exists.
00:27:48
It is a way for you to, it’s a low risk thing for the company. Let’s say you’re leaving a master’s degree, they can hire you as an analyst, keep you on for six months as an analyst and promote you to a data scientist. Low risk in the sense that you pay analyst about two-thirds of what you pay a data scientist. So, if you’re a data analyst making a hundred thousand dollars, an equivalent, data science will be making at least 150K. Probably more, but at least 150K. And so that’s why I think that transition exists. For the people who have master’s degrees it is an easy way for our company to judge someone’s business acumen with in a relatively low risk environment. And for the people who don’t have master’s degrees, it puts you in the basic space you need to be in order to show people that you have those data science skills and the job changes quite drastically.
00:28:38
Again, data analyst is not, it is a difficult job, but it is not a mentally extremely rigorous job versus being a data scientist. And the reason for that is because a lot of your work as a data analyst is getting the business requirements together and getting people together and coordinating people in order to actually pull together a Tableau dashboard or something. Or pull together a SQL where you need to actually show the data to an executive.
00:29:07
Versus a data scientist is like, okay, we do not know how much traffic is in our stores. I have this camera data, maybe I can use that. Then you spend a month looking into the camera data, find out it’s completely useless. Then you spend another month looking into another set of data, find out that that’s actually very useful. Oh, we have wifi in our stores. We can use that to estimate and model out how many people are in the stores at any given time. That’s a very mentally rigorous exercise. So, that’s why that transition exists. Now, why does the transition exist in data engineering? There’s a really common transition I’ve seen of data analysts to engineering-
Jon: 00:29:42
Data analyst engineering or data scientist engineering?
Shashank: 00:29:43
I’ve seen that too. Yeah, yeah, yeah, yeah. And that transition exists for multiple reasons. One, first let’s address the elephant in the room of the three major positions in the analytics world, data scientists, analysts and engineer. There are many more. They’re like ML Ops, DevOps. But these are the three really big ones that most companies will have. Analysts get paid the least. And as IC roles, individual contributor roles, they will continue to be paid the least on average. A data scientist and a data engineer of equivalent level will basically always make more money.
00:30:16
That’s the elephant to the room. Why do people transition? There you go. And I find a lot about corporate America, a lot about learning these new skills. It’s temporary pain glory forever. Because I’ve talked to analysts and they’re like, oh no, I don’t want to learn Python it’s too hard. But it’s like if you enjoy that temporary pain, eventually it becomes easy and then it’s just as easy as every other skill was for you. But now you’re making way more money. I see no loss here. I see no reason not to do it.
Jon: 00:30:46
Especially for people who enjoy that challenge.
Shashank: 00:30:48
Exactly. Yeah. Well even if you don’t enjoy the challenge, everyone enjoys more money. That’s what I don’t understand. I’ve had this conversation with a lot of analysts and I’m like, do you not want to make more? And I’m not talking like, oh, you’re making $500,000, let’s get you to 600,000. I’m talking like you’re making a hundred, let’s get you to 150 at a point where that difference can be pretty life changing. But yeah, I’ve noticed a transition from analytics into engineering because of people who get frustrated that they can’t get access to data as an analyst. And so they move into engineering to solve those issues. And I find that analysts who turn into engineers make some of the best engineers. Because they were the end customer at one day, not the end customer. They were the middle customer at one point in time. So, they understand the pains of analysts and what they need, what help they need.
Jon: 00:31:43
That is cool. That is a great perspective. I don’t think I’ve heard that perspective on the show before. But it makes perfect sense to me. What do you think about the Super Data Science podcast? Every episode I strive to create the best possible experience for you, and I’d love to hear how I’m doing at that. For a limited time we have a survey up at www.superdatascience.com/survey where you can provide me with your invaluable feedback on what you enjoy most about the show and critically about what we could be doing differently, what I could be improving for you. The survey will only take a few minutes of your time, but it could have a big impact on how I shape the show for you for years to come. So, now’s your chance. The survey is available at www.superdatascience.com/survey. That’s www.superdatascience.com/survey.
00:32:31
All right. So, you’ve talked about why people make transitions from data analysts to data scientists or to data engineer. But what are the key differences between those roles? I guess you’ve talked about that a bit. So, data analyst might not be writing code in a programming language like Python. So, maybe they’re writing SQL queries to extract data, but they’re typically not programming. I guess a data scientist will also typically be creating models. You mentioned that a bit. So, the data analyst is analyzing data, creating charts, summary data, understanding business requirements and being able to communicate those.
00:33:10
A data scientist can build a model that can take new data that it hasn’t seen before and make predictions. So, that’s probably a key difference between data analyst and data scientist. And then I guess the key distinction between data analyst and data engineer that you talked about is that the data engineer, again, because of those more developed programming skills, say in Python, is able to then extract data more themselves. Maybe they can be building data driven applications or applications that incorporate machine learning and be getting those into production, which is a big difference as a data analyst, you’re not programming production applications.
Shashank: 00:33:50
I would say that’s a pretty good summary of the differences. The way I like to think about it is in the pipeline of work, and no matter what I say, I want to make it very clear that all these positions are very, very necessary and they all drive a tremendous value. And any well-functioning analytics org will need all three of them, at least. In differing numbers there’s different theories as to how many engineers you should have per scientists, how many scientists you chat per analyst, stuff like that. But I like to see it this way. You start off with a data engineer, they go ahead and get the data ready for everyone else to use. And from there the data can go straight to a data analyst reporting on revenue, regulatory reporting, stuff like that can all be built by the analyst. If you want to know something that the data doesn’t directly say, that’s where you send the data to the data scientist and the data scientist can go ahead and build a model and say, okay, given this information we can assume or we can predict that XYZ is going to happen.
00:34:46
And then from there, usually what I would do is I would suggest that you give your data scientists work to a data analyst. This is not done enough in my opinion. But I think if you pair a data scientist and a data analyst together, you get a winning combo because the data scientist is so engulfed in the work that they’re doing, that presentation skills are not always top tier. But a data analyst has to have top tier presentation skills as part of the job.
00:35:08
And a data analyst gets regular practice presenting because their development cycles are usually shorter. It takes less time to build out Tableau dashboards than it does to build up machine learning models. And because of that, I think pairing a data scientist and a data analyst together where the data scientist builds the model and the data analyst builds out the charting mechanisms in order to actually monitor that model and make sure that people understand what’s going on is a winning combo. So, that’s another way I see it. There’s that chain of order where it goes from data engineer to analyst or data engineer to scientist to analyst.
Jon: 00:35:46
Nice. Yeah, that’s an insightful perspective that I hadn’t thought about before. But yeah, I can see how that data analyst, data scientist combo can be particularly effective in a business setting. Nice. All right, so we’ve talked about your transition into data engineer. Let’s dig into that a bit more. So, what do you do in your current role as a data engineer at Fanatics?
Shashank: 00:36:08
Yeah, so Fanatics is a parent org. I actually work in a company called BetFanatics, which is underneath it. It’s a sports betting startup. That’s part of the fanatics greater umbrella. So, as a data engineer, part of the reason I joined BetFanatics and one of the most exciting things about being here is that when I joined there was basically zero infrastructure. Nothing existed. And so it was a great opportunity to safely, because Fanatics is a minimum 27 billion dollar valued company. It’s privately owned, but minimum is 27 billion dollars. That is backing up this venture to get into BetFanatics or to get into sports betting. And so I know payroll won’t be a problem, I’ll be paid and everything. But I get to work on the ground floor with a group of extremely talented people.
00:36:54
So, what do I actually do? Right now it’s a lot of data modeling, building out what we think the data has to look like in order to satisfy regulators with our regulatory reporting and everything and what we think will be necessary for us to build out lines and stuff like that in the future. And making sure that we’re able to integrate our systems with any vendors who might be giving us data upstream. And setting up the rules and functions for our AWS implementation or Amazon Web Services, it’s Cloud implementation, to make sure that the data scientists and data analysts have an easy time putting together everything that they need in order to analyze their data.
Jon: 00:37:35
Well, sounds like a super cool role, Shashank. It’s amazing to have that kind of scenario where you don’t have to worry about getting paid, but you still get to work on a startup like project that has enormous potential to scale. It sounds really exciting. It does remind me of the scenario that I have with my company Nebula, and you don’t come across it very often. And so if listeners find a situation like that where you can be in a startup atmosphere or maybe even a completely independent subsidiary that’s being set up as part of a larger profitable holding company, that can be a really ideal situation for working in. Because you have so much upside potential without that risk of having to worry about whether you’re going to get a paycheck in the future. So, that’s awesome. So, I know that BetFanatics is specifically focused on sports betting, the broader company Fanatics is involved in the sports industry more generally. For our listeners, do you have any insight into how data are embedded in the sports industry today and how it could be transforming the sports industry of the future?
Shashank: 00:38:46
That’s a really interesting question. So, actually I’m able to answer this more because back when I was doing the whole two jobs thing during the early pandemic, one of my clients worked in the NBA. And it was interesting, all the data that they have available over there. So, there’s this thing called Second Spectrum, which is a series of cameras they have in stadiums, and I don’t know if they have it in all the NBA stadiums, but they haven’t been at least some of them. Like the Golden State Warriors for example. And basically it’s a series of cameras that can track every single player movement and can tell you exactly what play is being, is happening at any given time. As a bit of a side note, the funny thing is I feel like you could use that information to create a sportscaster that actually announces games and is able to … And I don’t think it would replace a human because there are [inaudible 00:39:37] there may be 10 sportscasters in the NBA. They’re highly talented individuals with great personalities.
00:39:42
There’s no developing an algorithm that could do that would cost millions of dollars. Why would you do that? Just pay these people, it’s way cheaper to just do it that way. But you can only imagine what you could do with the kind of data that says Player A swiveled 38 degrees in this direction and then passed the ball to player B, then player B went for a layup, player C came and shut him down.
00:40:08
A system that’s able to tell you all those plays has an incredible potential for people who are able to get access to that data. So, funny thing is I think that sports betting actually could get a lot more interesting if people able to get access to this kind of really hardcore data on how systems that work. One very interesting thing that I can think about is a lot of the sports books will have mid-game bets where, I don’t know, let’s say you’re playing baseball, will this specific player be struck out at this point in time? Which is a lot of fun. The logic of sports betting from, what is its benefit to society perspective is that it makes the game more enjoyable for the fans when … If you say, I bet you the cowboys are going to win and there’s nothing riding on that bet, if there are no stakes, it’s not interesting beyond just fandom for the team. Putting stakes on it or what makes it interesting, it’s why poker’s fun, but not when you play with no money.
00:41:11
So, I’m trying to connect how you have amazing data sets like a second spectrum that exist that I don’t believe people currently use in their betting algorithms. Now it might be because getting access to that data is, it’s usually gated access. But I think that you have the obvious examples of taking a lot more data, applying it to sports betting or even doing things like predicting player injuries might be a specific thing. Let’s say the NFL for example, concussions are a major problem in the NFL. What if there was a way to, based on the positioning of the players on the field, you could say this position leads to more injuries, leads to more concussions and more players. Maybe that should be something that we watch out for. Or I don’t know if concussions are one of those things that if they’re treated very early, people-
Jon: 00:42:04
Oh yeah. Oh for sure. Especially because if you don’t know the person has had a concussion and you put them back out on the field, they can get another one that makes the effects way more severe. So, so that kind of modeling to be able to say, based on the velocity and positioning of these players, there’s a high risk of a concussion here and they have to undergo some test before they can be put back up on the field.
Shashank: 00:42:26
Yeah. Or helmet sensors and stuff like that. So, I think when it comes to sports, a lot of the interesting data and a lot of the interesting effects that data will have on the industry are going to come through really cool hardware that’s being built out. Sensors that add no weight to a helmet, but it can detect the shock that a player is going through and communicate that immediately. You have the obvious example of Microsoft Services being used all over the NFL right now instead of pen and paper. How much that’s improved teams, I’m not exactly sure if that’s really done much. But it’s there. So, I think the opportunity exists and having worked tangentially to the industry with that contract I had with the NBA person, I can definitely say that there’s a lot of old dude thinking in sports. And a lot less data is being used than you think is being used.
00:43:20
And a lot of data is still pen and paper data, so it can’t be specific as to exactly what it is. But there’s a very popular event that’s held every year where all the results for the last whatever number of decades are taken in pen and paper. No one has turned this into a digital data set yet. It’s a really, really old man’s industry when it comes to data. People talk about the Golden State Warriors and all the data they use and everything, and I mean, sure they use some, but it’s not impressive in my opinion. So, there’s opportunities in there if you can get in.
Jon: 00:43:52
Yeah, interesting. Those are some great insights. I’m glad I asked the question and it sounds like there is a huge amount of opportunity to be digitizing information beyond what is currently being digitized, notwithstanding examples, the second spectrum example that you gave in the NBA. But yeah, still lots of opportunities in sports to be digitizing things and then even where we are collecting a lot of data, we’re just at the very beginning of what we could be doing with modeling and hopefully we can not only be coming up with better predictions about sports outcomes, but also hopefully doing things like avoiding serious injuries like concussions in the NFL. Very cool.
00:44:30
So, we’ve talked about your roles at the companies that you’ve worked at. Earlier in the episode you talked about FAANG companies, the big tech companies, which I guess today, even if you don’t like the Meta name, we could abbreviate them as like MAAMA. So, Microsoft, Amazon, Apple, Meta and Alphabet. And so doing data analysis at a company like Nordstrom that you were at previously relative to doing data analysis at a company like Google has got to be some big differences there.
00:45:09
So, you’ve spoken previously on other shows about the big divide in data maturity between data digital native companies like Google and those coming from older industries. And I guess to your point about these old man industries in sports, so those kinds of companies as well, they’re even further behind, way behind companies like Nordstrom. So, we’ve got this big divide between data digital native companies like Google and other companies from older industries for whom analytics is just a support function. So, can you elaborate on this big divide and what could be done to close that gap?
Shashank: 00:45:49
Show me your budget and I’ll show you what you are as a company. The big divide comes in budget more than, that’s where everything starts. At the end of the day, engineering is expensive. It’s not something that you can just spend a couple million dollars, you’re going to spend tens of millions of dollars on this. And the companies like Google, Apple, Amazon, these companies, Facebook and all, because it’s so core to what they do, they’re willing to spend just tons of money on engineers, on analysts, on scientists. And as such, I think because they spend so much money on engineers, they also just store a lot more of their data and formats that can be easily retrieved and allow analysts to easily go look into it and find out what’s going on. And find interesting correlations and potential causations that exist in that data.
00:46:47
So, I would say the biggest divide are, or a couple of things. It probably all comes down to budget initially. Everyone wants to hire the 10 x developer. The 10 x developer’s only going to go work for 200 plus a year though. And who has the money to hire someone for $200,000 plus per year plus stock. It’s going to be these really large FAANG companies. And the fact of the matter is their businesses are so profitable that they can afford to give their employees free lunch every single day of the week. Well, yeah, only a couple of them. But build these huge campuses and these extremely expensive parts of the country, pay engineers, if you’re talking Netflix, 500 plus per year in cash compensation. And the engineering salaries, they trickle down to everyone else as well. Data scientists get paid, I don’t know about 500, but data scientists get paid in the ballpark of that.
00:47:41
Analysts get paid less, but their pay is higher at companies like Netflix because they’re part of the tech award. So, yeah, I would say probably just comes down to budget and how much are companies willing to spend on this kind of stuff. So, yeah, in my opinion, budget is quite honestly where it all comes down to. How much are you willing to spend on this and what drives value in your org. I always like to tell people, if you’re going to ever take up a position, try and go to a company where your job is one of the most valued jobs out there. I have a lawyer friend and I asked him, is he ever interested in doing corporate law? And he says, not, or sorry, is he interested in doing in-house counsel? And he’s like, not really. In-house counsel, nothing wrong with it. But it’s not like the cutting edge of law is happening at law firms, companies. That’s what they do, they work in law. And the equivalent works with analysts being part of the tech org. You want to work at either a tech company or a tech first company.
00:48:43
So, there are three types of companies when it comes to, if you’re a tech person, tech companies, tech first companies and non-tech companies. Nordstrom, amazing company, it’s not a tech company. At its core revenue is generated from in person stores and that’s where the company started. Apple. Apple is 100% a tech company. All of their revenue is generated from selling technology and technology products. BetFanatics is what I call a tech first company. It’s primarily a sports betting company, but it’s a sports betting company with a heavy, heavy, heavy, heavy, heavy focus on being the industry leaders through a better tech product. And so those are the companies that, the tech first and tech companies that are willing to pay their analysts the most money. And as such, when you pay people more money, there’s generally going to be an increase in performance. You’re going to get better candidates, you’re going to get the top 10% of candidates, top 5% of candidates.
00:49:40
One of my friends works at Willis Towers Watson, amongst other things they do HR consulting. And he was kind of telling me, he puts together pay packages for people and he says, yeah, yeah, no, not every company’s looking for the top candidates. Why would they? Because if you want to hire the top candidates, you have to pay top dollar. Companies decide, okay, this function of our job, how much do we want to pay?
00:50:03
I was interviewing for a position at a major consulting firm and it would be a highly technical position. And what I found out is that this company amongst basically all the consulting companies, sorry, along with all the other consulting companies, they basically put all of their analysts and engineers in a warehouse in the middle of nowhere. And the consultants who are again, are the breadwinners for the company, they dump the work on the analysts and the engineers and then the engineers, analysts in their warehouse have to go finish it. You get paid decently well, but you’re not respected in the way that a consultant is. Don’t go work at McKenzie, Bain, Deloitte, unless you’re working as a consultant.
Jon: 00:50:43
Right, right. Yeah, yeah, yeah. That all makes perfect sense. Yeah, so a lot of really valuable tidbits there. A lot of great insight, Shashank, that you have into the marketplace for people working in data roles across a broad range of industries. And then crystal clear that it’s a budget that divides digital native companies from everyone else. And so yeah, if those other companies want to catch up, you’re going to have to pay. There’s no shortcut to getting that level of digitization and modeling and knowhow. It’s just money.
Shashank: 00:51:20
And I’d like to put it this way, engineering is new investment banking. Back in the 80s and to an extent the 90s, if you wanted to make a ton of money and you didn’t want to spend all a ton of time in school to become like a doctor, a lawyer or something, you become an investment banker. And even today you can make a lot of money being an investment banker. Today it’s software. If you work in software, those are the people that make just the most money in America without being in school forever. And even I’d argue if as an engineer or well yeah as an engineer or a data scientist, if you put in the amount of effort that medical school students put in, you’ll make more money than any doctor in America makes.
00:52:07
Because they don’t start making money until they’re like 30 plus. And then even then, it’s such a heavily regulated industry. There are only so many ways you can make money. You can have two jobs as an engineer. You can’t really have two jobs as a doctor because you have to physically be there for people.
Jon: 00:52:26
You could have relatively soon out of school, if you put the level of effort that somebody put into medical school, into your role as a data engineer, a data scientist, you could be making the kind of half million Netflix income that you’re describing while on the side getting your startup off the ground. That could mean by the time you’re in your mid 30s finishing your radiology residency or whatever and finally getting paid six figures and you’ve got all that medical school debt, you could instead be retired.
Shashank: 00:52:59
And that’s why I tell people, if you want to make money, don’t become a doctor. You need to become a doctor for the passion of medicine because software is a much more efficient way to do that. Yeah, I have friends that are in medical school and if you put in the amount of effort that they put into medical school, you would be a 10 x developer instantly. You would be the top 10% of developers easily.
Jon: 00:53:22
Yeah, agreed. All right, so you’ve had a lot of great insights into the industry. I’d love to hear if you have specific tool recommendations for our listeners that would maybe accelerate their journey towards being that 10 x developer, that 10 x data scientist.
Shashank: 00:53:41
Yeah. Okay. So, for the major tools, I work with Python, Tableau, and SQL. If you’re learning Tableau, I have a video on my YouTube channel on a Tableau, learn it. This is what you do. Easy enough. You can learn basically everything about Tableau and you can learn 90% of what you need to know in Tableau in about a weekend. When it comes to SQL, your select statements group by where window functions. If you know those. DQL Data Querying Language, then you’re basically good to go. For any interview. For Python, make sure you learn … So, Python, the way I like to describe it is data scientist and data analysts, or sorry, python’s like a tree and an engineer knows the trunk of the tree, the core of the language classes, functions, all that kind of stuff.
00:54:37
Data scientists and data analysts learn maybe a quarter of the trunk of the tree, and then they branch off into their own little world. That’s where you use Pandas, SKlearn. My brother’s like a software engineer. So, we’ve had discussions where we can read each other’s code, but we don’t actually know what’s going on because he knows all this really advanced core library stuff. But I know all these specific libraries, he has no idea what SKlearn is. Because I mean, why should he? It’s a very, very specific library, right?
Jon: 00:55:05
Yeah. For machine learning.
Shashank: 00:55:06
Exactly. Exactly. So, Pandas for data manipulation of tabular data, SKlearn for Machine learning, TensorFlow or PyTorch. You can use either for neural networks. I’d recommend doing that last though. But yeah, if you just know Pandas and SKlearn, then you’re basically solid as the basics of data science and data analytics.
Jon: 00:55:27
Nice. And do I remember from a call that you and I had earlier that the course that you used to teach yourself a lot of these Python skills was a super data science course by Kirill Eremenko, the old host?
Shashank: 00:55:39
Yes. [inaudible 00:55:41]. Yeah. The machine learning one, yes.
Jon: 00:55:42
Machine learning A to Z.
Shashank: 00:55:44
I highly, highly recommend it. I don’t do data science work today full time, but the algorithms I learned in that course, I still apply. I have a video coming out where I use K means, and I learned it from that course and Kirill puts it, he blazes through it. But in an appropriately fast way. He doesn’t dwell too much on theory, but enough to where you know what’s going on. Again, like I said, I’m not an academic, so I don’t care too much about the theory. I care about how this is practically used and Kirill’s course very much, yeah, that’s exactly what happens.
Jon: 00:56:19
Nice. All right, cool. Yeah, so those are similar to the recommendations that we had at the onset of this episode. So, Tableau, SQL, Python, but it was nice to have you dig into more detail into the Python libraries that you think are most important. So, Pandas, SKlearn, and either TensorFlow or PyTorch. Actually, I’ll also include, so we’ll have in the show notes, we’ll have your Tableau video that you mentioned. We’ll have the super data science machine learning A to Z Udemy course that Kirill made, and then I’ll also, I’ll provide a YouTube video on TensorFlow versus PyTorch and considerations that you might want to have as you consider adopting one or the other. Nice. All right. So, I understand from talking to you before the show that BetFanatics is doing a lot of hiring.
00:57:10
So, you made it sound like a pretty awesome place to work during this episode, given all the funding and the kind of startup atmosphere and the big scale of potential. So, you’re hiring data analysts, you’re hiring data scientists, you’re hiring data engineers, and you’re particularly looking for folks like that who have an interest in sports. So, Shashank I know that you do a lot of interviewing. What do you look for in the people that you want to hire for roles like that?
Shashank: 00:57:41
So, we’re looking for solid technical skills, first and foremost. There’s no two ways about it. If you want to work in this field, you have to have solid technical skills. That’s the most important thing. After that, we’re looking for the ability to work well as a team. We’re not looking for heroes. We’re not looking for 10 X devs who are in their cocoon for an entire week and then come out with a genius solution that no one understands. And we are looking for people who can really take on multiple roles and again, be compensated well to do that. Whenever anyone says take on multiple roles, make sure you’re being compensated well enough to justify that. As a bit of a side note, I’m a huge advocate of people being paid what they deserve to be paid. And I think at BetFanatics we’re quite good about that. So, we are-
Jon: 00:58:31
What do you mean by what kinds of multiple rules? What do you mean? So, they’d be doing data analytics and data engineering, for example. Is that what you mean?
Shashank: 00:58:39
I think you could drive a lot more value doing that. So, for example, if you go take a job at Amazon, you could spend 10 years in Amazon, never touch Tableau as a, actually, I don’t know if Tableau’s part of their stack. But never touch a BI tool as an engineer. Here, I mean you could. But it is really easy for you to drive a lot of value by building out a Tableau dashboard or seeing that, oh, maybe this code isn’t written as well as it could be written as an analyst and then pointing that out. Or as an analyst pointing out, hey, there’s a different AWS service we could be using for this exact same thing. In a startup the opportunities to do that are significantly higher.
Jon: 00:59:17
Nice. Yeah, sounds great. All right. So, solid technical skills. Work well on a team and be willing to take on multiple roles and then get compensated for that adaptability. Awesome. So, yeah, so look out for those opportunities listeners, if you’re looking for a great company, and particularly if you’re interested in sports. So, earlier this week at the time of recording, I asked my social media followers if they had questions for you. I told them that you were coming up to be interviewed and I didn’t exactly get questions. But I did get a lot of comments about you and some of which will spur discussion.
01:00:00
So, Richard, for example, who’s an associate data scientist at a company called HealthStream, he says that he loves your live streams, Shashank. So, that’s something that we haven’t even talked about yet on the program. So, maybe you could fill us in on those YouTube live streams a bit. And then Richard also asked to have you fill us in on your experience with the data crew, which I assume is going back to talking about Ken at the beginning of the episode. I think it’s kind of the crew that he’s put together for getting you guys physically together, a whole bunch of content creators in a house altogether and recording lots of content. I assume that’s what he’s talking about.
Shashank: 01:00:41
Yes. So, that was a very interesting experience to have. So, yeah, I guess kind of going into the livestream first, livestream 8:00 AM Pacific time every Thursday. I didn’t do one this week, but it happens basically every week. I just answer whatever questions people might have. As far as the data crew, that was a lot of fun. It’s a great opportunity to meet with not just data professionals, but people who also do stuff outside of their main jobs. 90% of people I meet … And honestly, this is the way it should be. 90% of the people I meet their day job is, that’s what they do. That’s all the productivity they have in a day from a data perspective. And not at all criticizing that is how the world should work. Most people should do their nine to five and then go live life after that.
01:01:29
But it’s really fun to meet people who also are out there making content, trying to help people teach, and it’s a group of friends that understand kind of the struggles of making content that I just don’t get in other situations. So, for me, that was kind of the coolest part about being a part of that. And again, they’re all just really cool people. It’s a self-selecting group of very hard workers. And it’s really funny because you’ll have people, Tina and stuff, constantly talk about how lazy she is and I’m like-
Jon: 01:01:57
Tina Huang.
Shashank: 01:01:58
Yeah, Tina Huang. I’m like, you have to be one of the hardest workers I know. You are, by definition extremely hardworking because you were doing your job and doing your YouTube channel at the same time.
Jon: 01:02:08
Yeah, it’s true. She is incredibly industrious and she does talk about how lazy she is. She does that in fact on Super Data Science, episode number 563, she’s our guest on that episode. I mean she talks about how, I guess what she’s kind of referring to there is that she has a tendency towards laziness, but as she talks about frequently on her YouTube channel, she has developed tricks, habits that spur her out of that innate laziness, I guess, and result in her being incredibly industrious.
Shashank: 01:02:42
And I mean there’s something to learn from that, right? I guess being industrious it’s doesn’t just come to everyone. Some people have to trick themselves into doing it. But as long as you get there, who cares how you got there? But yeah, it was a really cool experience. It was really trippy because, so I’ve never really been into celebrity culture myself. For me, my celebrities are YouTubers and MK, BHD, stuff like that. And one of those people was Ken Jee because I was watching … His videos, he very much has first mover advantage in this space in that he was one of the first people to come out with YouTube videos on data science. And I remember watching him early in my career, well, I guess I still am early in my career, but earlier in my career, and it was this trippy being friends with him and hanging out with him and just shooting the s*** with him about just life and everything instead of talking about data as well. So, yeah, no, it was an interesting trippy moment to get to meet your heroes.
Jon: 01:03:43
Nice. Yeah, is there a particular video that you recommend that we put in the show notes for listeners to watch about that time? Because Mark Moyou, who’s a senior data scientist at NVIDIA, he also commented that he loves the video that you did with Ken Jee for the datathon. So, Mark says he was laughing and entertained the whole time. So, is there a particular video that you could point us in the direction of?
Shashank: 01:04:07
It would probably be that one. So, there are two videos in relation to that.
Jon: 01:04:11
Yeah. What’s it called?
Shashank: 01:04:13
Oh, I think it’s called the Bright Data Iron Analyst Challenge. That was it. Yeah.
Jon: 01:04:17
Bright Data Iron Analyst Channel. Nice.
Shashank: 01:04:19
Yeah, so look up Iron Analyst Bright Data and you’ll see that. And then if you go to my channel, you’ll see me live streaming the entire event. So, basically, long story short, we’re given a random data set and we’re asked to analyze and build out a visualization of it and display it to a couple of judges. And I was the only one that live streamed myself doing it. So, you can see the entire process I was going through. And it might seem contrived, but it’s genuinely how the real corporate world works. It’s just squeezed down into two hours. So, you can see me go through the entire life cycle of building on a dashboard in two hours of like, okay, what do I prioritize? Where do I get this data from? What can I do? What can I not do? What’s realistic? What do the executives want to see? Stuff like that.
Jon: 01:04:59
Nice. Then, so that makes sense to the final comment that I’m going to read off of this social media post that I made about you, which is from Mark Freeman. So, he was also a guest on the podcast earlier this year. That was episode number 587, and Mark commented that he watched you build an entire interactive dashboard while live streaming in just a couple of hours. That’s must be what you’re just describing for this Bright Data Iron Analyst Challenge. And Mark says that you are next level in your data skills and your ability to communicate it.
01:05:31
So, given the experience that we’ve had and how well you’ve been communicated concepts in this episode, that doesn’t surprise me. But it’s nice to hear that he can also vouch for the technical abilities that you’ve been floating throughout this episode. So, yeah, so live streams on Shashank’s YouTube channel, you can see the master in action using his broad range of skills that he’s accumulated is a data analyst performing data science work, and now as a data engineer. All right, Shashank, it’s been awesome having you on the podcast. Before I wrap up episodes, I always ask for book recommendation. Do you have one for us?
Shashank: 01:06:09 Yes. My book recommendation is always Storytelling with Data for multiple reasons.
Jon: 01:06:15
[inaudible 01:06:17]?
Shashank: 01:06:18
Sorry.
Jon: 01:06:18
Is that Kate Strachnyi’s book?
Shashank: 01:06:20
It’s a Cole Nussbaumer Knaflic. Yeah, so [inaudible 01:06:24]. I Think KN, yeah, it’s Knaflic’s book, KN. So, really great book. Highly recommend it. And the reason is because it was the first book I read as an analyst, but the deeper I get into the technical skills, the more I feel like it’s important to pull myself back and understand that all data is storytelling. That’s the purpose of all of it. Why does any of it exist? It exists to automate stuff and to tell stories, and I believe that book does a great job communicating how do you do that effectively? And it’s a book that anyone can read if you do any kind of presentation at your job, which anyone of sufficient hierarchy in an organization will be doing presentations. It’s a good book to read.
Jon: 01:07:06
Nice. And yeah, prior to me becoming host of the Super Data Science podcast, so while Kirill Eremenko whom we mentioned as the creator of that machine learning A to Z course, while he was still host of Super Data Science. He hosted Cole Nussbaumer Knaflic, the author of that book that you just mentioned, Storytelling with Data. And so that’s episode number 395, so listeners can check that out as well. It’s a great recommendation. Thank you, Shashank. All right, so obviously if people want to keep in touch with you after this episode, one great way to do that is to be following your YouTube channel, which we’ll have links to in the show notes, including watching your live streams and interacting with you there. Any other social media handles that listeners should be aware of for tracking your work and staying up to your latest?
Shashank: 01:07:51
Probably LinkedIn is a good one. You can follow me on LinkedIn and I’ll try and post there a little bit more often. That’s probably the way to go though.
Jon: 01:07:57
Nice. All right. Shashank, thank you so much for being on the Super Data Science Podcast. Thank you for getting up early to make it happen with us. It’s been so much fun having you on the show. You’re a delight to chat with, and yeah, look forward to catching up with you again in the future.
Shashank: 01:08:13
I had a great time. Thank you for having me.
Jon: 01:08:20
What a jolly time I had with Shashank filming today’s episode. In it, Shashank filled us in on how you can learn Tableau in a weekend and it will accelerate your capacity to provide business intelligence reporting as well as potentially be enough to land you an entry-level data analyst role. How Tableau, SQL, and Excel are the must-have hard skills for a senior data analyst position. While Python is a hugely valuable, nice to have. How senior data analysts shine at requirements gathering, project management, presentation skills prioritization, and their willingness to learn. How data analysts who become data engineers can be some of the best data engineers out there because they appreciate the needs of the folks who will be analyzing the data downstream and how data manipulation with Pandas machine learning, with SKlearn and deep learning with either TensorFlow or PyTorch are the critical Python library based skills for data scientists.
01:09:13
As always, you can get all the show notes including the transcript for this episode, the video recording, any materials mentioned on the show, the URLs for Shashank’s YouTube and LinkedIn, as well as my own social media profiles at www.superdatascience.com/623. That’s www.superdatascience.com/623. If you enjoyed this episode, I’d greatly appreciate it if you’d left a review on your favorite podcasting app or on the Super Data Science YouTube channel, and of course subscribe if you haven’t already. I also encourage you to let me know your thoughts on this episode directly by following me on LinkedIn or Twitter and then tagging me in a post about it. Your feedback is invaluable for helping us shape future episodes of the show. And if you’d like to engage with me in person as opposed to just through social media, I’d love to meet you this week at the Open Data Science Conference West.
01:09:59
That’s ODSC West, it’s running in San Francisco from today through November 3rd. I’ll be doing an official book signing for my book, Deep Learning Illustrated, and we’ll be filming a super data science episode, live on the big stage with the world leading deep learning and cryptography researcher Professor Dawn Song as my guest. In addition to those formal events, I’ll also just be hanging around grabbing beers and chatting with folks. It’d be so fun to see you there. All right, thanks to my colleagues at Nebula for supporting me while I create content like this Super Data Science episode for you. And thanks of course to Ivana, Mario, Natalie, Serg, Sylvia, Zara, and Kirill on the Super Data Science team for producing another terrific episode for us today.
01:10:37
For enabling this super team to create this free podcast for you, we are deeply grateful to our sponsors. Please consider supporting the show by checking out our sponsors’ links, which you can find in the show notes. And if you yourself are interested in sponsoring an episode, you can find our contact details in the show notes as well as by making your way to jonkrohn.com/podcast. Last but not least, thanks to you for listening all the way to the end of the show. Until next time, my friend, keep on rocking it out there and I’m looking forward to enjoying another round of the Super Data Science Podcast with you very soon.
