101: Expert Data Analytics Panel Tells All (100th Episode Celebration)

[00:00:00] There's nothing actually wrong with you. You're not the problem. Often the system is the problem and so it's okay like if you mess up with certain styles of interviews, you're not good at certain things, then maybe you'll be a better fit at other places. If you enjoy it, generally the jobs...if I got the job, I found the interview fairly easy. Like it felt much easier and I ended up being really good fit for those places. Welcome to the Data Career Podcast, the podcast that helps aspiring data professionals land their next data job.

[00:00:30] Here's your host Avery Smith. We made it to 100 episodes you guys! This is the 100th episode of Data Career Podcast, if you're listening, thank you so much for listening and supporting the show. I can't believe we got to 100 episodes. I think it's something like 90% of podcasts don't get past episode 7. So thank you guys for listening and giving me someone to talk to. I really appreciate it. I think you're going to enjoy this episode because it's a little bit unique. It's a panel interview that I

[00:01:00] did with some of your favorite data career professionals, data content creators. I did it in person in Charleston and North South Carolina. I Charlotte, North Carolina, Charleston, South Carolina. It's just a little confusing but yes, South Carolina, Charleston. I got to meet some of the coolest people, nicest people in the Data Content Creation Space in person for like a four days retreat work session, whatever you want to call it. And it was a lot of fun. So I got to hang out with Ken G.

[00:01:30] Alex the analyst, Sundas Khalid, Jess Ramos, Megan Liu and a bunch of other really awesome data content creators that you've probably seen on LinkedIn or YouTube or Instagram or TikTok.

[00:01:42] Yeah, it was a lot of fun to give you a little bit behind the scenes. Basically got there Thursday, kind of went to a big group dinner, hang out all together. We were all staying in the same houses so we spent a lot of time just socializing talking, getting to know one another like in person versus just...

[00:01:58] We've been... you've known each other for years online but never really met in person type of a thing. On the Friday, we got to hear from the sponsor of the event which is Posit. So thank you Posit for bringing us together. They're the creators of our studio, the tidy verse and a bunch of other R and Python tools that you may have not used yet but you probably will use in the future and get to learn more about their products and also participate in a hackathon.

[00:02:23] I was like trying really hard to win the hackathon and I lost to Shishankalanathy again, he beat me last time and he beat me again. We built like the coolest tool and our demonstration it didn't work. It failed for some reason. Actually we think it failed because of an HDMI cable that we plugged in which I know sounds crazy but it was pretty involved in anyways so I lost the hackathon, dang it. But it was fun trying to code something up. My teammates were Jack Blandon and Keith Galley.

[00:02:52] And we built something that I think I was pretty proud of. It's a lot of work to code underneath the pressure but we built some pretty cool stuff.

[00:02:59] Then on Saturday, I did a bunch of podcast recordings so you'll hear those upcoming in the future episodes. I did one with Alex the Analyst which will be coming out next week.

[00:03:10] I did one with Kenji Josh Starmer from StackQuest on YouTube.

[00:03:14] To this panel podcast that you're about to listen to was some awesome guests. We had Elijah Butler because I've seen him he's a senior data analyst at Humana and he's pretty big content creator on TikTok.

[00:03:25] But we had Rashad Knives Becker, creator, he's like a VP of data at MetLife, creator on LinkedIn.

[00:03:33] We had Monica K. Royale, also another LinkedIn creator we had Kenji.

[00:03:38] So like we had the whole crew you guys it was a lot of fun. I think you're really going to enjoy this episode.

[00:03:43] It's kind of like a little bit different because it is a panel episode where I basically ask them a few different questions and you get to hear some of their backgrounds and experiences and how they're the same and how they're different.

[00:03:55] Some of the resources they recommend on your data journey. So I think it's a fun episode to celebrate 100 wreaking episodes you guys. I can't believe it. Pretty insane.

[00:04:05] The other thing I want to tell you guys is if you're listening this in March on March 25th, I'm launching something called Interview Week.

[00:04:12] It's going to be a bunch of fun free events to help you guys feel more confident in your interviews.

[00:04:19] So if you are interviewing or you will be interviewing shortly, you guys got to check out this event.

[00:04:24] I'm just stay tuned on my email, my newsletter and my LinkedIn probably for updates for that for right now.

[00:04:30] But I'm pretty stoked to about that. And that's in preparation for the software I've been building the last year.

[00:04:36] It's called Interview Simulator. It's finally going to be ready to come out.

[00:04:40] I've literally been working on it for a year. It's time to release it to the world and give it a go.

[00:04:47] So that's coming out. It's really cool. It's basically interactive video interview prep that is going to give you coaching and tips on how you could possibly do better in your interviews.

[00:05:00] So that is coming out as well. And I'm super excited about that.

[00:05:04] So stay tuned. You can learn more information, interview simulator simulator.io. But I'll be talking about it in the upcoming episodes and stuff like that.

[00:05:13] We have a lot of fun for your resources with that. And I'm pretty stoked to be launching that.

[00:05:17] Okay, without further ado, let's get into this week's episode and this panel episode for the 100th episode.

[00:05:25] Once again, thank you guys so much for listening. I appreciate each and every one of you.

[00:05:29] And if you haven't left a review or a rating on Apple or Spotify podcasts, please do so. Hopefully I've been able to provide you value or two over the last 100 episodes.

[00:05:39] And in exchange for that value, all I ask is you leave a review on this review or a rating on Apple or Spotify.

[00:05:45] Yeah, if you have done it, thank you so much. If you haven't, it literally takes like 35 seconds and it really helps the podcast find more people just like you and expands our reach, which allows us to continue making great content for you guys. That's free.

[00:05:58] Alright, let's get into this episode.

[00:06:00] Alright, welcome to the 100th episode of the Data Career Podcast.

[00:06:08] Woo!

[00:06:10] Good job on that.

[00:06:12] Thank you. Thank you for listening and we got some awesome guests here to kind of share who they are and what they're doing.

[00:06:20] We're here in South Carolina at Creator Meetup. Thank you, Posit for bringing us here and we got some great people on this panel that I'm going to let them introduce themselves a little bit about who they are and what they do. So we'll start with you.

[00:06:34] Yeah, it's so much so, you have a specker.

[00:06:37] I'm a VP Data Science, I co-lead the global team at MetLife for Data Science and lots of leadership strategy challenges with AI, right?

[00:06:46] Which is our giant opportunity but could be a giant threat. Also, I just want to say all of these are my own.

[00:06:52] And yeah, excited beer.

[00:06:55] My name is Elijah Butler. By day, I am a Data Analyst by night. I am a Data Analyst. Mostly TikTok and I do some stuff on the other things too but I focus on making content that helps people get their first job in data analytics and then just talk about everything that I believe makes a really good data analyst and I'm very blessed and honored and humbled and queasy to be here.

[00:07:20] I am Monica K. Royal, I'm with Nerd Nurseman, I am Data Career Strategist and Instructor helping people transition from their current jobs into a fun and exciting data job.

[00:07:34] And fun fact, I love to dance.

[00:07:38] I also love to dance. My name is Ken. I am a head of Data Science at Sports Analytics Consulting firm. I also do Data Science content and podcasting on Ken's Nurse Neighbor's Avery.

[00:07:54] Congratulations on getting 200.

[00:07:56] Thank you, I appreciate it. Ken's got a great podcast, you guys haven't checked it out. Just launched a new one too.

[00:08:02] What's that one called? I forgot the name. It's called the Exponential Athlete. And it's got some great sports stories on there. Yeah, a little bit different outside the data realm but just hyper focusing on what makes the greatest athletes of all time. Great by reading a lot of books.

[00:08:16] I love it. What's really interesting is I'm not sure, I think I know all of your guys' backgrounds. What did you guys do before data? Did you guys do stuff before data?

[00:08:26] Okay, I applied golf and management consulting. And management consulting? What about you Monica? What did you do?

[00:08:31] I was an auditor so a lot of security, cybersecurity, IT stuff.

[00:08:37] Oh, basically I just went to college. I'm 25. When I got out I didn't know for sure I would go into data but I pretty quickly got into data once I got out of college.

[00:08:49] You were an economic student. Yeah as an economic student I got some experience in college doing like forecasting

[00:08:55] and regression analysis. And once you do that, SQL becomes a lot easier than like learning to do regressions I guess.

[00:09:06] I did a four-month stint in inside sales for learning and development company. Before that I wanted to get a PhD neuroscience and then I realized both those paths were terrible for me.

[00:09:17] So then I was figuring out what to do at home and I thought data science, that sounds cool. Also money.

[00:09:23] That's true. Data is a lucrative career. Okay now we're going to open it up, feel free to answer any question that you guys would want here.

[00:09:33] But I'd love to hear, what did you guys' favorite data tool or technique? This could be like regression, it could be like you love Excel. I know many of you guys do.

[00:09:43] What's your go-to favorite tool or technique? So for me I don't know if it's purely a data science tool or technique but something that I found unbelievably useful is monocarlo simulation.

[00:09:55] To me that is a great way to make very complex variables tell a story or be able to simulate what things might happen in the future.

[00:10:08] And I think it's very often overlooked because it doesn't fall into the traditional training and then testing like a regular model does.

[00:10:18] So I've used that approach for a lot of my own work and I would encourage people to learn about monocarlo simulation because it can be very useful for a lot of problems that don't necessarily super easily fall into the data science lifecycle or data science pipeline.

[00:10:36] In particular things that are closed loop where you can only have, if you have two dependent variables on something, if one goes up the other one might have to go down. And in most machine learning models you predict those independently.

[00:10:50] Wow that's a great one.

[00:10:53] I'm actually going to say that my favorite tool is your brain just specifically like thinking outside of the box, like extremely outside of the box and bringing your specific perspectives from say your past life, your hobbies and all of that.

[00:11:11] When you bring different perspectives to a certain data problem, you're able to solve a lot deeper and involved problems also I'm old school and really love Excel.

[00:11:23] That's great that's awesome.

[00:11:25] I'm really partial to some unsupervised learning techniques like all the different clustering methods.

[00:11:32] I like them because they often require a human in the loop to interpret and so they can give you insights about the problem you're working on rather than treating it as purely a prediction problem.

[00:11:44] They still feel a bit like magic to me and because there are many, many ways to group like things together which is what clustering is.

[00:11:54] And they also lend themselves well to visualization so they can make some interesting and impressive visuals for stakeholders.

[00:12:02] So they're useful from a breaking the problem down perspective, they're useful from a visualization and talking to the stakeholders perspective and they're cool.

[00:12:12] That's a great answer. I love that.

[00:12:14] The clustering can make some really interesting data visualizations at the end.

[00:12:20] That's a great answer to everybody. Let's do this one. We're kind of doing a random group of questions here but do you guys feel threatened by AI?

[00:12:28] No, I think AI is a great opportunity. So from my perspective, I'm not hands on every day anymore like I used to be and so I look at data science with a very broad enterprise perspective but with a lot of empathy as a technical person and as a data scientist.

[00:12:46] I think being a data scientist before Gen AI exploded on the scene was really challenging for many reasons. One of those was getting stakeholders to listen to you and to care about what you're doing.

[00:12:58] So I've personally had the experience of trying to just find someone who would give me the time of day to do some kind of project with like starting a data science practice from scratch.

[00:13:10] That was challenging and also quite luck-oriented, luck-based. You couldn't control those variables, right?

[00:13:17] And that is the context in which you do all your work. So with AI now there's this explosion of interest and now suddenly we're back on the roadmap again.

[00:13:26] So people want to listen to what we have to say and we as data scientists can provide advice and thought leadership to our business stakeholders to help them understand what's out there, separate hype from reality and help them make better decisions.

[00:13:41] It's really I think the generational opportunity for math and stats oriented people to become to be viewed as equal partners with the business.

[00:13:51] And so definitely I'm not threatened at all.

[00:13:54] I will take the other side of that. I am very threatened. Not in terms of my work necessarily.

[00:13:59] I think the unique skill sets that I have or that other people who are doing hands-on data science work have are not replaceable in the short term by AI.

[00:14:10] But I think more broadly if we think about AI and how fast it's moving, that makes me feel quite threatened.

[00:14:16] You think about the Turing test. I think it was in the 40s or 50s or 60s when Alan Turing probably should know a little bit closer to the date but when he created that and he thought in is either 10 or 20 years, that that would be a solved problem.

[00:14:31] And it took almost 50, 60 years for that to happen.

[00:14:34] And now a lot of the AI advancement is happening dramatically faster than we would predict.

[00:14:40] And this is a trend across all of AI is that we historically have underestimated how long things would take and now we're dramatically overestimating how long things would take.

[00:14:51] And that suggests to me that this AI boom is an exponential curve because as you imagine, exponential would mean that at the beginning it starts to failo and then there's a hockey stick.

[00:15:01] And so if we view it in that light, I think humans have a very difficult time understanding exponential growth and the implications of that.

[00:15:10] And at a broad level, that is scary to me.

[00:15:13] It's because we're dealing with a system that is growing so fast that humans cannot conceptualize that.

[00:15:19] You look at social media, you look at all these things that are unbelievably addictive already.

[00:15:23] And that is because they're hacking our brains in ways that we have not been able to adapt to evolutionarily.

[00:15:30] So I think that there are some dangers with AI in terms of how fast it's moving.

[00:15:34] I don't know exactly what that looks like, but I think that in that frame, which is different from your frame, there's good reason to be threatened and skeptical and wary of the things to come.

[00:15:45] That doesn't mean you should be scared, doesn't mean you should necessarily become a plumber and hide under rocks.

[00:15:50] But it means that we should all look at these things with a grain of salt and maybe take some precautionary measures or more precautionary measures as we go.

[00:16:00] I'm going to say yes and no actually, so no because AI is not going to take your jobs.

[00:16:07] There's always going to be a human element to the situation.

[00:16:11] But yes, because of the fact that humans are involved and we don't know exactly what will be built or where that will progress to.

[00:16:21] Am I threatened by AI? I'm going to answer this the same way that I do when someone asks if I'm threatened by sharks only when I'm at the beach.

[00:16:31] Okay, great. We're kind of close to the ocean right now. Are you threatened by the sharks in there or?

[00:16:37] And AI.

[00:16:39] Okay, both of them are kind of making you uneasy.

[00:16:43] I swear whenever sharks start learning about large language models, we're all screwed.

[00:16:48] Oh yeah, we're done for when that happens.

[00:16:52] What kind of a kid is talking about? You got to constantly be learning and learning with AI now.

[00:16:58] What's one of your guys' favorite resources for data learning, specifically data learning?

[00:17:03] What do you guys like to use? I mean, not to fluff his ego anymore but my number one resource is probably the Alexey Analyst YouTube Channel.

[00:17:12] Especially when you're going for like your first job in data analytics, you can learn the basics and more about all of the skills you need to learn.

[00:17:20] You can... he has road maps, he has videos on resumes so there's something for everything that you need as you're trying to become a data analyst.

[00:17:29] I just talked about how threatened I was by AI but I think ChatGPT, Claude and any of these types of things.

[00:17:35] They're great resources for learning. I also... I think that they're great resources for learning because of their flaws.

[00:17:41] So we know they can hallucinate, we know that there are challenges with the things that they produce.

[00:17:46] There's variability to them, there's a lot of these things.

[00:17:49] And I actually think that that creates a very good learning environment.

[00:17:52] If I'm using one of these tools to learn or to help me coach me through a problem, I know that there's a chance that it's not right.

[00:18:00] So if I comment with that frame, it helps me to do the complementary or supplementary research to make sure it truly is right.

[00:18:07] It also gives you immediate feedback, it's a dynamic problem or a dynamic solution which I think is exactly what almost all of us need to produce the best results from learning.

[00:18:18] So it's just that this immediate rapid-fire feedback being able to ask or prompt it to ask us the right questions,

[00:18:24] to get and to help lead us to the solutions that are valuable.

[00:18:28] I will also say that it only is useful if you use it the correct way.

[00:18:33] So not just asking it for answers but asking it to help teach you something.

[00:18:37] I think that is by far one of the best long-term use cases of AI is personalized education for almost everyone,

[00:18:44] because I think education is a one-size-fits-all approach currently.

[00:18:48] And to me this is one of the single best resources to help even globally solve a problem of a mismatch between the student level and the education.

[00:19:01] I can't pick a favorite, I love them all, they all have the things that they bring to the table.

[00:19:08] So I'm just going to say in general the internet and find your own style and the way that you like to learn.

[00:19:16] There's a saying you might have heard no plan survives first contact with the enemy.

[00:19:22] I actually think the best place to learn to be an analyst is your stakeholders, the actual people you're serving.

[00:19:28] Meaning that you can learn all the technical tools and you can supplement your knowledge with endless courses and videos and the like

[00:19:37] and they are good for getting started.

[00:19:39] You can also build projects which will guide your learning with an outcome.

[00:19:43] But the ultimate outcome of a data analyst is essentially to be a really useful partner

[00:19:48] to give information and insights that other people will act on and find useful because they don't act on it then what's the point.

[00:19:55] So the stakeholder is actually the best long term teacher because you will make mistakes, mistakes that are not really taught in courses.

[00:20:04] You will see how they respond to the things that you do and if you're able to internalize that well and use their feedback to improve

[00:20:13] not just the specific thing you did but the craft and how you approach problems in the first place.

[00:20:18] That's the only real path to being a really successful analyst in the long term.

[00:20:23] So my answer is reality.

[00:20:26] Reality, I like that.

[00:20:28] I think for me, none of you guys mentioned books.

[00:20:30] I like some books.

[00:20:31] Yeah, Ken reads like nobody.

[00:20:33] In fact, I was up this morning at like 830 and Ken's over here talking about this book and that book and you were like

[00:20:41] you're all these 10 reads a lot of books.

[00:20:43] So I think for me, I don't know if I have like a favorite data analytics or data science book but I just interview Josh

[00:20:51] Sharmer.

[00:20:52] I love the illustrated guide for machine learning by StackRus.

[00:20:56] Really good.

[00:20:58] I love storytelling with data for data visualization by Cole Mathleck and then have any of you guys read any of the Edward Tufty books?

[00:21:06] The Edward Tufty books are like my favorite books ever but they're, anyways, they're pretty in-depth data visualization books.

[00:21:14] So I like those serious books.

[00:21:16] I got a little bit on that.

[00:21:17] I have two books that I really like in data.

[00:21:20] The first one is small data.

[00:21:22] I read it before getting into the field and it gave me a lot of respect for the value of observation,

[00:21:28] really, really careful observation.

[00:21:30] Instead of thinking that if you just throw more data at the problem then it will solve it.

[00:21:34] And I think the industry at that time in the mid teens was moving really in the throws of big data obsession.

[00:21:41] If we just throw more information then the machine will find the patterns and it'll all be good.

[00:21:46] And I think large language models are now exposing the challenges of that, especially with safety, ethics and the like.

[00:21:51] If you just throw a lot of messy data, you're going to have to do a lot in order to correct for that.

[00:21:56] All sorts of architecture and complexity and money spent in order to solve those problems that would be solved with clean better data.

[00:22:05] So one is small data observation.

[00:22:07] The other was my start, the only book I read before I got my first job which was Introduction is Statistical Learning with Applications in R, Hastings in Tib Shirani.

[00:22:15] That was like my legend book.

[00:22:17] I don't know, I just love it because it's visual enough but they also have really good word explanations for what's happening.

[00:22:24] So if you feel uncomfortable staring at lots of math symbols, I found it like a pretty good balance point.

[00:22:29] Yeah, I think they have that online for free nowadays too.

[00:22:33] Yeah, which is quite nice.

[00:22:34] I apologize, I have a fire I have to put out.

[00:22:36] Oh no problem.

[00:22:37] This has been amazing, Avery.

[00:22:38] Again congrats on 100 and now you can use the title Why Can You Walked Out of my 100th episode?

[00:22:46] Ah, perfect!

[00:22:47] I appreciate it.

[00:22:49] Are we still doing books?

[00:22:50] Yeah go ahead.

[00:22:51] Sure, yeah.

[00:22:52] So I really like weapons of math destruction by Kathy O'Neill I believe it is.

[00:22:58] It's a really cool look into how biases are in machine learning algorithms.

[00:23:04] So one example is with underwriting processes so when you apply for a loan or when you apply for a credit card they look at certain things like your gender,

[00:23:13] your credit score and so on and so like those biases get introduced in the algorithm and it's a really, really interesting dive into that.

[00:23:22] Have you guys read it?

[00:23:23] I have not.

[00:23:24] I need to check it out.

[00:23:25] Check it out.

[00:23:26] The last question I wanted to ask you guys was what's like if you guys were in an interview or maybe you're conducting an interview.

[00:23:34] What's an interview tip that you'd want to give someone?

[00:23:37] My biggest thing and I am a very unqualified person to give any advice on this.

[00:23:43] I've never interviewed anybody but my thing that I really try to do is to make some kind of connection and just be friendly like obviously if you're going to work in data or any field

[00:23:54] you're going to be expected to know certain skills and the like but everyone that they're interviewing for the most part is going to have that.

[00:24:01] They're going to want to hire somebody that they want to work with so like if you can kind of start almost a friendship,

[00:24:08] like it's still an interview but like try to find some common ground, ride the vibes and the like and I really think just making any kind of connection during an interview can only help.

[00:24:18] I second that bring your enthusiasm, bring your true self to the interview and also to practice interviews I like to share to think about recording yourself and then watching it

[00:24:30] and then you can see like how you speak, how you explain things and it really does help like improve the way that you present yourself during an interview.

[00:24:41] Pretend the interviewer is your friend and that you're actually already colleagues and you're just working together to solve the problem rather than them judging you.

[00:24:49] Firstly it's kind of true because they're going to be your colleague if you get hired and then you know why not practice that before you join?

[00:24:58] It'll make the nerves go away, that's one thing.

[00:25:02] Another thing is in the larger context I think a lot of interview practices are actually kind of dumb.

[00:25:09] I thought this when I first got into the field but I didn't necessarily trust my own judgment.

[00:25:14] I thought wow this is really stressful for no reason you know and they would do certain things especially live coding in front of someone who's watching you.

[00:25:22] I would always cave and like lose all my knowledge when faced with interviews like that.

[00:25:27] And so if you encounter intense interview practices and you think you know this is one really hard and two it's not necessarily like a real fact similarly of the job meaning if you're doing like lead code challenges but the actual job,

[00:25:42] you're not going to be doing much of that. Like actually trust your judgment there's nothing actually wrong with you.

[00:25:48] You're not the problem often the system is the problem and so it's okay like if you mess up with certain styles of interviews,

[00:25:55] you're not good at certain things then maybe you'll be a better fit at other places.

[00:25:59] If you enjoy generally the jobs, if I got the job I found the interview fairly easy like it felt much easier and I ended up being really good fit for those places.

[00:26:09] So if you found interview really hard and then you didn't get it that's alright that place wasn't the right fit.

[00:26:15] So believe in yourself like you don't have to you don't have to fit like everyone else's box, you don't have to fit the prestige box.

[00:26:23] I would fit your box and then I would define your own box over time.

[00:26:27] All right well that was our 100th episode panel and you guys can follow all these people who have all their links in the description down below.

[00:26:34] Thank you guys for being on the show.

[00:26:36] I appreciate it thank you for having us.

[00:26:38] Yeah, I think you should build a box and then you should stand on top of your box and then break the box and build a better box but thank you for having me and congrats on 100 episodes.

[00:26:49] I love that so much.

[00:26:51] Yes, thank you congratulations for your 100th episode and yeah.

[00:26:56] All right hopefully you guys enjoyed that episode.

[00:27:03] I know I always love hearing from multiple people with the same questions and just hearing different ways to approach different things, different backgrounds and stuff like that.

[00:27:12] Thank you guys for listening to all 100 episodes of the Data Career Podcast.

[00:27:15] We have interview simulator week coming out really soon.

[00:27:19] You guys can check the my LinkedIn and email for that but if you guys need interview prep this free week is going to be really useful.

[00:27:25] For helping you kick off and get improved in all your interview needs.

[00:27:31] So be on the lookout for that and I got nothing else for you so I will see you guys later.