105: Do You Have to Be Good at Statistics to Be a Data Analyst w/ StatQuest Josh Starmer PhD

[00:00:00] Do you see the math as scary in statistics?

[00:00:03] Yes, it's totally scary. I hate math. Math hates me. Yeah, math is scary and math is really hard.

[00:00:09] People who are scared of math, they can be like, oh, well, Josh is also scared of the math still

[00:00:13] and he's like the king of statistics on YouTube. I've got one last thing for you.

[00:00:17] Oh, let's go ahead. Triple BAM. That's what I was hoping we were gonna end on, was a triple BAM.

[00:00:23] Welcome to the Data Career Podcast, the podcast that helps aspiring data professionals

[00:00:27] land their next data job. Here's your host Avery Smith. BAM! Welcome back to the Data Career Podcast.

[00:00:37] I'm here with Josh Starmer, PhD, better known as StackQuest right here on YouTube.

[00:00:44] BAM. The famous BAM. If you guys haven't checked out Josh before or StackQuest,

[00:00:50] you guys gotta check it out on YouTube. Basically, in my opinion, this is the...

[00:00:55] You've made so many contributions to the statistics and data analytics and data science

[00:01:00] community with all these statistics YouTube videos that you've made. First off, thank you.

[00:01:05] My pleasure. Second off though, really Josh's role in society is he's basically the YouTube

[00:01:14] that you turn on when you have a midterm in college for your stats class. That's basically

[00:01:18] what Josh's life, life work is. He's the person you turn on when you have a stats midterm in the

[00:01:26] middle of the semester. How does it make you feel? It makes me feel pretty special because it could

[00:01:32] have been anyone and it ended up being me. Well, we appreciate it. We're going to talk about Josh's

[00:01:38] background and how he got into statistics. We're going to get into a little bit about

[00:01:43] what you need to know with statistics in this episode. Be ready. Here we go.

[00:01:48] All right. My first question, we're going to get to your background here in a second,

[00:01:51] but I want to start with... I have a lot of new people who are transitioning teachers or their

[00:01:57] warehouse workers and they want to become data analysts. These are people that maybe got an

[00:02:03] English degree or a teaching degree or maybe they didn't even go to college and they want

[00:02:08] to become data analysts. A lot of the times they're scared about math. Do you see the

[00:02:15] math as scary in statistics? Yes, really? Yeah, it's totally scary. I hate math. Math hates me.

[00:02:22] I don't work well with symbols and I don't want to offend anyone in the

[00:02:28] wonderfully beautiful country of Greece, but when I start seeing Greek characters on a page,

[00:02:34] my hands start to sweat. Yeah, math is very, for me, very hard and very intimidating and

[00:02:46] I've been working with it my whole life and I still, when I see equations,

[00:02:51] I'm just talking about it. I get sweaty problems right now. Yeah, math is scary. Yeah, math is

[00:02:56] scary and math is really hard. It is for me at least. Yeah, super hard. That's so interesting.

[00:03:02] Even though you make all these YouTube videos, you work for an AI company.

[00:03:07] You're still a little bit scared. Yeah, exactly. It's a little bit like

[00:03:12] I guess mountain climbing or whatever. I see the equation and it intimidates me. It's so big

[00:03:17] and tall but part of me also wants to understand it because I know what it does. It's this

[00:03:24] real compact thing but it's somehow very powerful and so I want to understand why it's

[00:03:29] powerful and what gives it its power. I guess what gives it its power and what the big deal is.

[00:03:38] When I see an equation, I guess I get that fear of missing out thing. There's something encoded

[00:03:43] in there that I don't get and everyone else is like, check it out. I get it. I understand. I'm

[00:03:48] like, what? But I don't get it. That's another thing. Yeah, so I want to,

[00:03:55] I know it can do cool things. I just want to understand it. So I will persevere and

[00:03:58] I'll climb that mountain. I don't care how tall it is and I'll get to the top and I'll be honest,

[00:04:03] this is the other thing that even though these things scare the dickings out of me,

[00:04:08] when I get to the top of that mountain, and I solve that, I'd not solve the equation but

[00:04:14] I figure out what the equation means and represents, it's one of the best feelings ever.

[00:04:20] I mean, I love, I just get such a rush and it is like climbing a mountain or

[00:04:24] accomplishing some other goal that was difficult and you had to persevere and you finally got it.

[00:04:30] And you know how proud, proud of yourself you are when you finally did it. That's me when I've

[00:04:34] finally understood an equation. That's so interesting because now Josh and I met,

[00:04:38] I think twice now, kind of on these prolonged little trips. We should also mention that thank

[00:04:43] you to deposit for bringing us here together. Yeah. But I've learned that, you know,

[00:04:48] Josh, if you guys watch his videos, you know that he loves math. You know he loves to

[00:04:52] sing and play music. Josh also loves to cook. I do. And so I feel like all three of those things,

[00:04:57] there's things you were probably just like really curious about and you're like, I want to understand

[00:05:01] how does that sound get made or how does this amazing steak get cooked? Yeah, it's true.

[00:05:09] There's a lot of curiosity in what drives me is curiosity almost for everything.

[00:05:14] That's so interesting. I taste something and go, how did this happen? I hear something and I go,

[00:05:19] how did that happen? And I want to figure out how to recreate it or make it myself and

[00:05:24] kind of break it down into the pieces and just understand. Makes sense. And really what I was

[00:05:30] asking is the math scary problem. I tell people the math isn't scary. It's not scary. That's

[00:05:36] what I say. But no, no, no, no, that's good. Say your true opinion. But what my point here is,

[00:05:41] is I think it's actually awesome that like people who are scared of the math,

[00:05:45] they can be like, Oh, well, Josh is also scared of the math still. And he's like the king of statistics

[00:05:50] on YouTube. So it's like, is statistics scary? Yes. But Josh did it. And if Josh did it, I can do it.

[00:05:57] Yeah, I hope that inspires some people because it's very true. It's not about

[00:06:03] looking at an equation going, Oh, that looks easy. It's not about that. It's about

[00:06:08] or that looks hard or something like that. It's everyone can understand these things.

[00:06:12] And even the most complicated equation really is just little pieces of it, relatively simple

[00:06:18] things combined, right? You can have a really dense equation, but it's just addition and

[00:06:24] multiplication. And we all we all learn how to do that in second grade or third grade or

[00:06:29] maybe fourth grade. I don't know when I learned it. It was a long time ago. But you know what

[00:06:32] I'm saying? Like, when you break it into the smaller pieces, you can get it down to something

[00:06:37] where you're you can it's it's understandable. And then and then I start playing around with it.

[00:06:44] And sooner or later, I start seeing what it's really doing once I, you know, I can see the

[00:06:49] individual pieces doing their thing. And that's something you're really good at in your videos.

[00:06:54] They're all very like modular almost like furniture where it's like you could switch it

[00:06:59] like you learn one thing and applies here and applies there. And you do that as well

[00:07:02] in your book. Yeah, is it the illustrated guide for machine learning? Is that the name of it?

[00:07:06] It is the stat quest illustrated guide to machine learning. Yeah, if you guys haven't checked that

[00:07:12] out, we'll have a link to it down below. If you're just if you're just starting on the

[00:07:16] data world, machine learning might be a little bit much. But if you're like ready to learn

[00:07:20] machine learning, I think I wish I would have started with your book because you do

[00:07:24] first off, you do two things really well. Number one is you do what you just said,

[00:07:27] where you make it like very modularized one bit at a time where it's like,

[00:07:31] I'm not just taking machine learning a whole. You like help help me start like things I

[00:07:34] kind of recognize from stats glass maybe like a like a p value. We'll get to that in a little bit.

[00:07:39] And then also you make it really fun. Yeah, it has like a bunch of cartoon drawings and

[00:07:43] and honestly, I think you changed the format of the book right now. Yeah, it used to be

[00:07:47] a landscape. Yeah, a portrait. So now it's more normal but the original landscape. I know it was

[00:07:52] a it was a piece of art that I mean I'm sure the new ones a piece of art too but

[00:07:56] it like it looked like a good like belong in a pop art museum. Yeah, I was really pleased

[00:08:01] with the original version but unfortunately I couldn't get that one printed internationally.

[00:08:06] So I had in order to get the book more places, I had to reformat the whole thing. It took like

[00:08:14] you were doing that last time. I was in the basement of the house where everyone was and I

[00:08:18] was just like like a machine just going crazy just reformatting every single page.

[00:08:24] But I did it. You did it. Yeah. And it's out. Yeah. Anyways, I love the book. I think

[00:08:28] it's really great if you're just getting into machine learning. Honestly, I think this is

[00:08:32] something you need and if you're like me where it's like I've been doing machine learning for

[00:08:36] since I don't know what 2016 it's almost eight years now. But like I also am not like in the

[00:08:45] trenches of machine learning every day. Like yesterday we had a hackathon and that was like

[00:08:50] my first time coding hardcore in like a couple months. Yeah. And I was like oh yeah, this

[00:08:54] is kind of get on your bike anyways. The books really good for that for me because it's

[00:08:57] like oh yeah, wait, what are the rules of this and I can always go back and check the book and

[00:09:01] anyways love the book. Let's talk about like for people who are just getting into statistics.

[00:09:07] What are like three like if you're just learning statistics, you're just getting started.

[00:09:12] What are like three things that someone should know? Let's start with one.

[00:09:16] This is going to sound weird and it actually really applies to anything not just the statistics

[00:09:20] but learn how you learn like we were just talking about how scared I am of equations.

[00:09:26] When I took statistics, they just put a bunch of equations on the board or the wall or the you know

[00:09:32] and they just said this is the equation for variance. This is the equation for the mean. This is

[00:09:36] the equation for this thing called a t test and there was just a bunch of equations

[00:09:40] and none of it made any sense at all to me. Now does that mean I'm never going to be a good

[00:09:45] statistician? Does that mean I should just pick another job career or class or something like

[00:09:51] that? And the answer is no because it just meant that looking at a bunch of equations isn't the

[00:09:57] way I learn. I'm a real visual learner and I did not learn statistics during that class at all and

[00:10:04] I got a really bad grade but five years later, I mean I knew I had to learn statistics for my

[00:10:10] job and I knew I had to figure it out so I just started trying to use pictures and I drew

[00:10:17] pictures about what was going on and I tried to illustrate what the statistic, what the equation

[00:10:24] was doing and when I did that I got it. And as a result I became quite an accomplished

[00:10:33] statistician and one that's you know I mean I feel like I'm bragging but I was you know highly

[00:10:40] respected. People came to me for all kinds of problems because what they discovered is

[00:10:49] I didn't just know the equations, I actually knew what was happening. I knew what was really going

[00:10:53] on and I'd learned it in an incredibly deep way and it was all just because I found another way

[00:11:03] to learn. So I think for any topic but especially statistics, I personally had struggled with that

[00:11:12] one because to me it's a weird topic. It's unlike anything you're taught in grade school

[00:11:21] for example like when you take math in grade school you know your high school calculus or

[00:11:29] all those things there's one answer and that answer is like just one number.

[00:11:38] And so our whole lives up to a certain point is all about solving for single individual numbers

[00:11:47] and statistics is using those same mathematical techniques, addition, multiplication, not that

[00:11:54] complicated stuff but the results are weird in that it's not just a number, it's a mean plus

[00:12:02] variation. What does that mean and why can't I just use the mean all the time? Why isn't that good

[00:12:10] enough because it used to be like when you're in high school they're like calculate the average

[00:12:14] and you calculate the average and you get a number and that was good enough then why isn't

[00:12:19] it good enough now? So I had a lot of conceptual difficulty wrapping my brain around the fact that

[00:12:25] I was using math but I couldn't, I wasn't, the old things that I was doing weren't good enough

[00:12:33] all of a sudden and I didn't understand why and I didn't understand anything that was going on

[00:12:37] because it was just it was all things I'd done before but in a way that made no sense and so

[00:12:45] it was very hard for me so I had to, so I'm just saying like for statistics was when I was a profound

[00:12:50] thing where I had to find a new way to learn that wasn't just by looking at equations and

[00:12:56] you know doing whatever we did in the class because that didn't work for me and so I

[00:13:01] think that's what it's important for anyone. It was super important for me so when you're

[00:13:06] starting out with statistics and you're like I hate this I don't get it doesn't make any sense

[00:13:11] maybe it's because it's you're learning the wrong way or that you're being taught the wrong

[00:13:14] way and it's not the teacher's fault everyone of these different some people are going to respond

[00:13:19] positively to the way that teacher's teaching so it's kind of up to you or it's up to me to

[00:13:25] find a way to make it work on my own I mean and one thing that's nice is there are lots of youtube

[00:13:30] channels there's mine obviously but there's lots of others that have different approaches so you

[00:13:35] can try mine and go you know it doesn't work for me and you can try someone else's and you can

[00:13:39] find someone that does resonate with you and so it's not like you know I'm throwing you out

[00:13:46] and making you have to learn all this stuff on your own there there's teachers out there you

[00:13:50] just have to find the one and it may not be the first one you find but hopefully you'll find it

[00:13:54] sooner or later. Yeah figuring out what actually works for you in your learning style is so

[00:13:59] important because actually I'm usually like in college uh I pretty much played a golf game

[00:14:06] on my phone while I was there because I don't learn from like a powerpoint presentation so like I

[00:14:13] go to class and I try to listen I don't get bored I would just play golf on my phone yeah and then

[00:14:17] I'd actually learn by doing the homework yeah it's like okay I go do the homework okay now I need

[00:14:22] now okay now I understand why kind of need to do this yeah I'd go like in the book read okay

[00:14:27] that makes more sense for me and usually I don't really like theory very much um which is what

[00:14:33] you kind of do in your videos. That's what I do yeah a lot of theory which which is awesome

[00:14:36] because if you want theory and theory works for you you know watch Josh I don't really cover

[00:14:39] statistics that much but I'm like very practical yeah like most of my youtube videos are like

[00:14:43] this is how you do this data analysis with this data in this situation it's like

[00:14:49] way too practical um but the cool thing is is you know some people are going to resonate with

[00:14:53] with yours yeah some people are going to resume and that's that's great you gotta figure out what

[00:14:57] works best exactly yeah okay so that's the first thing figure out how you learn yeah the next

[00:15:01] thing if you figured out what you learn yeah what should you learn next and statistic the next

[00:15:05] thing is try to wrap your brain around the concept of variation and variance the core the whole reason

[00:15:12] for statistics to begin with is that is variance and variation um you know when I was talking about

[00:15:24] you know when we learn how to calculate the mean in high school they didn't then say

[00:15:27] and now calculate the variance of it you know and so when I finally had to start doing that

[00:15:32] I was like what is going on why can't I just have I don't why don't I just report one number

[00:15:39] why do I also talk about variance but what what statistics is is every time we do something every

[00:15:45] time we see something or experience something every time I go to my favorite restaurant and order

[00:15:50] french fries I get a different number of french fries I don't or different sizes or different

[00:15:56] um you know there there's just there's it's never the same order of french fries

[00:16:02] you know 10 times in a row it's never exactly the same right and so what do you what do we

[00:16:09] do about that can I use that variation as a way to make informed decisions for example

[00:16:17] where am I going to get the most french fries most often like there might be some restaurant

[00:16:23] like occasionally gives me tons of french fries but it's rare and usually they only give me a little

[00:16:28] bit and there might be another place that's like all over the place all the time and we can use that

[00:16:34] variation we statistics what it does is it takes that variation and quantifies it and once it's

[00:16:40] quantified we can use it to make informed decisions and that's the power of statistics in

[00:16:45] and nutshell it's all about quantifying understanding and using variation to make informed decisions

[00:16:53] very good definition I like that definition you can see that josh is a very he's got these concepts

[00:16:59] defined in his head yeah very well I like that because yeah like basically statistics does exist

[00:17:05] for variation and trying to figure out because we as humans we we know that like okay this is

[00:17:10] more french fries than than this french fry but like you can't remember last weeks or five

[00:17:15] months ago's french fry size or the number of french fries and so really what that allows us to do is

[00:17:20] is take take the variation or like the differences in occurrences in life and put it into a numerical

[00:17:26] form and we as humans can be like oh 10 is bigger than 5 so then we can make like you said

[00:17:32] informed decisions so yeah I really like that definition right all right good bam that's a

[00:17:37] bam yeah right there okay so we got we got learn how to learn yeah we got understand variation

[00:17:42] what's the third thing the third one is um uh I don't know exactly what the right word for it is but

[00:17:49] what I I call it experiment with whatever the the thing you're learning if you have a computer

[00:17:56] if you got excel you can you can like if I learn a new statistical concept so there's this thing in

[00:18:03] statistics called the central limit theorem and you know and if you look at the equation for

[00:18:08] the central limit theorem it's really complicated that makes me yawn just thinking about it yeah

[00:18:13] it's like it's this horrible intimidating thing um but what it means what that equation means is that

[00:18:20] and this is going to sound weird when we calculate averages for something say like the average number

[00:18:24] of french fries we'll say like I calculate the average of the friend for french fries at the

[00:18:30] restaurants I go to and you go to the same restaurants like a day later whatever and you

[00:18:35] calculate the average on the days you go to and then maybe somebody else goes the next day and

[00:18:40] they calculate the average you know say say like we got like 10 people and we all calculated

[00:18:45] our averages well if we plotted a histogram of those averages it would have a a belt shape

[00:18:52] curve it would it would be approximately what we call approximately normal uh in statistics and

[00:18:59] um but doing all that work you know getting my 10 friends to eat french fries 10 days in a row

[00:19:07] and then do and repeat that 10 times or something like that getting everyone so they everyone has

[00:19:13] 10 measurements that they can use to calculate their mean that takes really time consuming

[00:19:17] yeah a lot of french fries it's a lot of french fries and I do love french fries but

[00:19:20] I may not want to wait 10 days before I get my next order in you know what I'm saying

[00:19:24] so what's cool is I can I can simulate that on a computer I can I can write a program to like

[00:19:30] generate a random number of french fries and and do it 10 times and then calculate the mean

[00:19:36] and then repeat that process 10 times so I've got 10 means and then I can do it in excel and I

[00:19:41] can plot a histogram and I can see it and then there's this huge complicated equation

[00:19:46] that just boils down to this nice uh histogram that looks like a like a like a curve and then

[00:19:53] I can uh and I can say well what happens if um if we all order small fries

[00:20:02] does it change so instead of you know the random number of french fries being like a

[00:20:08] reasonable number of french fries like 20 or 30 maybe we only get five or on average

[00:20:13] you know what happens to the to that histogram does it gets wider does it get narrower

[00:20:18] what happens if we order super size fries so we get like a hundred fries with each order

[00:20:23] does that change the shape of the curve I can all I can do all this stuff on a computer and I can

[00:20:28] play with it I can see what happens and the more I play with it the more I understand what's going

[00:20:34] on and and then when someone goes well we're only we're only collecting this much data uh but

[00:20:40] we're gonna assume it's normally distributed because you have to do that for certain statistical

[00:20:44] tests I'll look at them and go well that's going to be a kind of a rough approximation

[00:20:49] or they'll say we're getting all this data and be like and I'm and I'll know because of those

[00:20:53] experiments that I did that that's going to be a really good approximation you know so um so

[00:20:59] that's a way that I've I've learned statistics it's just by playing with it and experimenting

[00:21:04] on my computer in excel or r or whatever tool you want to use um that's how I've learned

[00:21:12] a lot about statistics playing around playing around having a good time thinking about french

[00:21:16] fries practice yeah it does help if you if you make your statistics more french fries or donuts or

[00:21:22] something yeah exactly yeah if you're not a savory guy you can do donuts yeah uh maybe donut bites

[00:21:27] how many don't eat your holes yeah um that's so interesting so the learn how to learn yeah

[00:21:34] the second one is um understand variance or try really hard to understand variance and

[00:21:40] to give you a tip on understanding variance I I always draw things I try to illustrate what

[00:21:45] variance means in this context you know uh and so that's that's something you can okay

[00:21:50] I would do it visually and then yeah then experiment that's so interesting yeah so when I

[00:21:57] originally wrote this question yeah I thought the answer was going to be p values hypothesis

[00:22:02] testing and regression no but I think that just goes to show how like practical I am and

[00:22:08] like I'm like but anyways that's so interesting that that like really in the day what you're saying

[00:22:14] is the most important thing about statistics is like spend time with statistics like like practice

[00:22:20] it's like no one actually really understands it the first time they hear it and you get

[00:22:25] better at it as you go yeah it's a weird it's a weird field uh and it's and it doesn't make a

[00:22:30] lot of sense to a lot of people because they haven't really spent a lot of time playing with it

[00:22:38] I think and all of a sudden you're like okay I get it um you know play with it that's that's so

[00:22:43] interesting I want to um I want to kind of change change gears oh actually let's talk about this though

[00:22:48] um memorization yes this is something that this is like a question that I have in general like do

[00:22:56] you have like all of like stats memorized no should you no no I'm a main idea person

[00:23:06] to be honest I am very bad with details uh and especially memorization um

[00:23:15] but and so I don't think it's that important I've got a terrible memory I forget all the time

[00:23:20] and so I mean there's certain things you need to remember which which is like main ideas like

[00:23:26] what variance represents right and I feel like you need to have these real those concepts like

[00:23:32] really well down really well but in terms of like what's the equation for doing a linear regression

[00:23:39] I you know I might not even know that off the top of my head um t-tests um you know which is a

[00:23:48] t-test is a sort of a fundamental statistical test I could not tell you what the equation is

[00:23:53] for that you know most things I couldn't tell you the equation for the the concepts are important

[00:23:59] to to quote-unquote have memories or at least understand what what they are yeah and one of

[00:24:04] the reasons why is like you never actually use the equations yourself right you plug it into excel

[00:24:10] and it gives you the answer and what's important is you know how to interpret that answer rather than

[00:24:15] like oh I know exactly how it's you know it's doing this and it's dividing by this and it has

[00:24:19] this thing called degrees of freedom and blah blah I don't you know you don't have to remember

[00:24:24] those things you what's important is that you've got your data you put it in your spreadsheet

[00:24:28] you put it in R or whatever you're using you did the test here's the output what doesn't mean

[00:24:34] who cares how it was done you just want to know what it means so yeah the concepts and interpreting

[00:24:39] the results which is actually interesting because um I was I just interviewed Alex the analyst one

[00:24:44] of the things that we talked about that makes a good analyst is like less your technical skills

[00:24:50] and more your ability to explain what you did technically and how it relates to the business

[00:24:56] kind of the statistics it's the same it's like it's not like do you have

[00:25:00] you know like a like a t-test formula memorized it's like can you know the situation where you'd

[00:25:06] want to use a t-test yeah have a computer use it and then being able to interpret the results

[00:25:11] exactly yeah so there's some memory involved you got to understand that there's times when this

[00:25:17] test is useful and there's times that it isn't and you might have to you have to remember

[00:25:22] those little details um but in terms of the nitty gritty stuff you don't have to remember any of that

[00:25:29] okay very neat I want to go kind of like back into time okay um back to when you were in school

[00:25:35] you got a music degree I do but then you also got a computer science degree I did yeah which is

[00:25:40] interesting because a lot of people I actually like um I've helped a few musicians who have music

[00:25:45] degrees become data analysts um I've helped a music therapist I don't know if you ever heard

[00:25:49] of that oh yeah yeah became a data analyst um Aaron Sheena I think episode like 57 or something like

[00:25:54] that um but it's so interesting because a lot of them never have the computer science degree so you

[00:25:59] you did both yeah as an undergrad yeah and then you went to biostatistics is that right yeah yeah I've

[00:26:05] uh I have a phd in something called bioinformatics bioinformatics yeah okay it's biostatistics

[00:26:11] so you've always been kind of curious about computers and math and stuff like that

[00:26:16] yeah I mean when I was a kid I was really just drawn to computers um

[00:26:23] I mean I like to play outside too and climb trees that was also fun uh but I've all you know I've

[00:26:28] always felt some affinity towards computers and I don't really know why um um I don't know but

[00:26:36] but but um I don't know like when I was a kid I used to try to make video game my own video

[00:26:41] games and things like that I just thought was fun uh to do and so when I went to college I loved music

[00:26:50] and I loved computers and so I did both and um it turned out that uh that I have actually had

[00:27:00] the opportunity to do both uh professionally and it turns out I don't like doing either one without

[00:27:06] the other um you know if I'm coding all day I love going home and playing music all night yeah

[00:27:14] and if I'm playing music all day you know I you know I used to be in a touring rock and roll band

[00:27:20] and do some things like that that's fine but when I go home I want to code and do math all night or

[00:27:25] whatever you know like work on on data science all night I just love I love the I love it

[00:27:32] all yeah and so it's it's been I feel very fortunate that I've been able to pursue both

[00:27:39] that's it's very neat to do both and if you haven't watched Josh's YouTube you'll you'll

[00:27:44] go watch one and you'll understand how he uses music in a very silly unprofessional way

[00:27:49] and it's awesome it's very awesome um but so so anyways you're you're you end up working in

[00:27:56] a lab yeah and um you have this you know bioinformatics phd yeah and so everyone's kind of asking you

[00:28:02] hey Josh I got this I'm doing the stats thing you're like how do I do this yeah and the people just

[00:28:08] keep coming and asking you kind of the same question and one day you get sick of it yeah

[00:28:11] and you're like I'm gonna make a video yeah explains it that's kind of how it happened uh

[00:28:17] I mean it was partly that like I don't like I think all computer programmers don't like doing

[00:28:22] the same thing more than once yes you know they like the variety and and so I think I had a

[00:28:30] an impetus or motive to teach the people around me how to do their own stats I wanted them to

[00:28:36] present with calm and I wanted them to understand what they were doing and how they were doing it

[00:28:40] so that when if someone asked them a tricky tricky question they could they could you

[00:28:44] respond correctly um and so yeah I started making these these videos well what I started off doing

[00:28:51] that I started off teaching people the concepts and they were basic really crummy PowerPoint presentations

[00:29:01] not fancy looking at all but I would do these things and like at the time I called them stat chats

[00:29:06] Friday morning stat chats um but it was an academic lab and so uh I started having to

[00:29:14] repeat the lectures because there'd be new people in the lab and again I don't really like

[00:29:18] doing the same thing more than once and so I um I was like I got an idea I'm gonna put the videos

[00:29:25] that these presentations I'll put them up as YouTube videos and it solved all kinds of problems

[00:29:31] not just the having to repeat myself but one thing that's really cool about YouTube that I

[00:29:38] is it meant that the people in the lab could learn when they needed to learn

[00:29:44] rather than when I had time to teach right um and what's important about that is like I could

[00:29:51] teach them some statistical concepts I could say hey on Friday we're gonna talk about variants

[00:29:56] and they'd be like great uh and they would come and they would learn about variants

[00:30:00] and then they'd go maybe a year and a half before they actually need to do anything right and by

[00:30:06] then they've forgotten and maybe gotten it confused and it was like that whole that time was just

[00:30:11] wasted and learning takes work they spent I mean it takes work to teach and it takes work to learn

[00:30:19] your brain has to change that's work and and so they did we all did work for no reason at all

[00:30:26] because by the time they needed it they'd forgotten and have to relearn and so what was what was

[00:30:31] cool about YouTube is I would just put it up there and they'd be like well I'm reading this

[00:30:37] manuscript and it says it's got this thing called r squared and I don't know what the heck that is but

[00:30:42] I'm gonna check Josh's channel and look he's got a video on it I can learn it when I need it

[00:30:50] and when I'm using it and so that was like a real big plus for YouTube and it wasn't about

[00:30:58] and so anyways my early videos are all I used to work in a mouse genetics lab so all my videos

[00:31:03] were designed just for my co-workers they're all the examples are like imagine you have a mouse

[00:31:08] that weighs this much and it's this big blah blah blah you know so they're all mouse genetics

[00:31:14] examples and I'm talking about genes because these are geneticists and it's funny a lot of people

[00:31:18] watch my videos now and they're like these these examples are terrible you know they're like

[00:31:24] why are you talking about mouse genetics I'm not a mouse geneticist but I the purpose was

[00:31:30] was not to have videos that everybody watched I was really just trying to help my co-workers

[00:31:34] specifically and that's something that is I think was also very important maybe to my success was

[00:31:43] I knew who my audience was yeah I wasn't just being like I think I can teach r squared here's

[00:31:49] what I'm gonna do it my my my attitude was like I'm trying to teach r squared to these

[00:31:55] specific people how what's the best way for me to do that and the best way for me to do that for

[00:32:01] them was to use mouse examples because that's what they they could relate to that a lot of statistics

[00:32:06] for me is hard to when I'm learning statistics one thing that's really hard is is because it can

[00:32:12] be applied in so many different contexts and this probably bothers you too they use the most

[00:32:16] abstract examples ever they're like imagine you have an urn filled with two different types

[00:32:22] of marbles and I'm like yeah I'm like why should I do that you know so I'm like okay you've got these

[00:32:29] mice and so immediately my co-workers they wouldn't have to imagine something that's kind of abstract

[00:32:36] they could imagine something that they work with every day yeah and they could relate to it

[00:32:41] and that I think that was a big part was knowing who I was I was trying to communicate to

[00:32:46] and not just being like oh ever this is for everybody um and being really general about it I

[00:32:52] tried to be real specific and give lots of of specific details so that the people that watch

[00:32:57] it could could relate for me that's that's always really useful because like even now when I think

[00:33:03] about doing um like some more complex machine learning stuff like like PCA and um like clustering

[00:33:09] and and classification stuff like that I think back to when I was working and I was using it

[00:33:15] when I originally learned it like those examples and I always think through

[00:33:19] all of the actual math and those examples and then I always try to relate it to what I'm doing now

[00:33:23] okay yeah that's how I did it then what's this part what's anyways so that is really useful and

[00:33:27] one thing I just want to highlight is from you you were as a like bio statistician uh

[00:33:33] bio was a bio informatic or something like that um you your job wasn't to do that necessarily

[00:33:41] but you were documenting your learnings and like your whole life looks different because you

[00:33:46] documented your learnings that's true yeah one thing that um they actually said one of the

[00:33:50] posts that people said was every they said every data scientist is a content creator

[00:33:55] and I thought it was kind of weird at first but now when I look back at it it's like any

[00:33:59] anyone who's doing anything technical you're always going to get benefit out of documenting

[00:34:04] your learnings and sharing with others yeah and like your life would look a lot different

[00:34:09] if you'd never had done that that's exactly right which is crazy because it's hard to know in the

[00:34:14] moment it's scary I mean maybe at the time you're just putting yourself on YouTube for them not for

[00:34:18] other people to see but anytime you're making any sort of videos your voice or your face could be

[00:34:23] on there and it's scary and stuff like that and I just think there's so much to be benefited by

[00:34:28] you know sharing your learnings as you go yep your whole life is different that's why it's

[00:34:32] called stat quest yeah is it's my it started out as my quest to learn statistics it was my

[00:34:38] adventure yeah and you weren't even I think you told me this last time yeah like sometimes you'd

[00:34:43] you'd make the videos to understand the concepts yeah exactly yeah like it was just you learning

[00:34:48] in public exactly yeah that's how it works is I um I learned in public in front of everybody

[00:34:55] with a lot of people watching which is scary yeah which is scary but there's so much to be

[00:35:00] benefited from it and that's one of the things I try to work with my students is like just

[00:35:03] post on LinkedIn what you're learning yeah and like you'd be amazed what what doors can be can be

[00:35:08] open yeah you can build communities that way and and make friends and and help people too because

[00:35:14] because someone might say hey you're learning that uh I'm learning it too but I I couldn't

[00:35:19] figure this part out or or you know it's like it's it's oh yeah it's great it's really great

[00:35:25] because I think people I think I think although learning is work and hard I think it's also

[00:35:30] fun and I and it's a fun activity and it's fun to do with other people even if it's

[00:35:35] remotely and things like that better than doing it by yourself yeah always a lot of time all right

[00:35:40] Josh yeah that's um that's all we had for today can we end with one more thing uh I've got one last

[00:35:45] thing for you let's go ahead triple bam that's what I was hoping we were gonna end on was a triple

[00:35:50] bam so if you guys don't follow Josh make sure you uh go check well this is linked to his

[00:35:55] YouTube down below and uh we'll have a link to his book you guys can check that out I I

[00:36:00] before I met Josh I bought a copy so I'm not biased and saying it's a great book right and even last

[00:36:05] time last time you had a few copies that I think you gave to people and I I was like I bought mine

[00:36:10] I have an original it's signed by Josh now so go check it out guys all right thanks Josh thank you