How I'd Learn Data Science In 2023 (If I Could Restart) | A Beginner's Roadmap

if you're watching this video you might be in the position I was in not too long ago thinking about getting into this field of data science but feeling a little bit intimidated at this Prospect of having to know maths coding and if you don't have a background in this stuff wondering if you could ever break into this field listen I completely get it I went through that same situation myself and now I'm a data scientist with an actual job and everything the whole process took me about two years but honestly that could have been cut down by at least a year maybe even a year and a half to make it take six months and even though I'm at the beginning of my career and I can't promise to get you into Fang or anything like that I think perhaps I could help a couple people get from knowing nothing to getting one step closer to getting a job in just a few months so here's how I would learn data science in 2023 if I had to start from scratch and just a heads up I did go to university for data science but for this video I'm gonna assume you want to go down the self-taught route honestly if I had to restart I still would go to UNI just because it makes it easier to find some jobs where they just want you to be able to take that box that you went to UNI but like I said let's assume you're going self-taught because you can still make it that way the first step is just to learn what data science is okay don't skip this step I know it might seem like a no-brainer but just taking the time to really understand what data science is and how it's used by businesses is really important trust me I made the mistake of just jumping straight into it thinking I knew what data science was big mistake don't assume that data science is just about coding and maths and then just jump straight into learning that stuff okay take time it doesn't have to be a long time maybe a week maybe two weeks to make sure you concretely understand what data science is why businesses care about data science and how it's used in different contexts and domains it might even be worth just going on to YouTube and looking up a really long tutorial and treating that as a walkthrough about what your day-to-day life would look like where you have to build projects to provide value because nowadays with stuff like chat GPT On The Rise just knowing how to code isn't enough you have to know what you're doing when to implement a certain style of code to get a certain result and why are you doing that so again don't jump straight into coding take time to understand what data science is okay step two now that we know what data science is let's jump straight into coding data science is such a broad field that in some jobs you'll be coding literally every minute of every day basically and others you only be doing that once in a while but you have to know what coding is and how to do it to some extent at least that's why it's important to have a baseline language now most people say it doesn't matter whether you know r or python first and most jobs will say you just have to know one of the other but if you had to learn one it's python let's be real python is the more used one across the industry that's just a fact and python is such a broad programming language I find it funny when people are like oh go Learn Python it's like what parts of python am I going to use as a data scientist could you be specific please and honestly the education system gets a lot of flack but University was actually really useful because they taught us the basics first the basics of coding and it made it so much easier to build on top of that so this is exactly what I would learn and when so when you're first learning forget about libraries and knowing how to build a model or anything like that just learn how python works as a programming language the best ways to start is download something like vs code and you can explore python in there and especially if your laptop is not that powerful you can use something like Google collab instead which is free and really powerful and just take time to explore the language okay so here's a list on screen of what I would focus on when learning Python and roughly in what order I'll do it and trust me the first thing I would learn is data types because if you're anything like me most of your errors when you start off with will be because you're using the wrong data type and if you're learning in 2023 you're in luck because now you have chat GPT back in my day we had to look on stack Overflow to find our errors but as much as possible please as a beginner don't overly rely on chat GPT use it to teach you how to code not just wax stuff in chat GPT and assume you've learned from that make sure you're asking it to explain to you what was done and why and also just so you know it's not always correct so just be careful of that and after learning data types move on to assigning variables lists dictionaries working with data frames and then how to do simple procedures like for Loops while Loops as well as defining functions and how to put in different arguments and defaults in there if right now you're like oh this sounds really complicated trust me it's not you're just hearing new terms so it might sound a little bit difficult but once you get started you can do this stuff trust me and at this stage when you're just learning you really only need two libraries pandas and numpy pandas to work with data frames and what have you load in your csvs and numpy to do mathematical procedures and maybe a visualization Library like matplotlib but I prefer plotly because it's a little bit sexier but really you can do so much with just those three libraries now that we know this Baseline level of python the next thing is to do a project and here wait this really come over here right we're up close and personal okay I would just kind of listen to other people saying do projects but I wouldn't actually do the projects because when you're learning tutorials the Temptation is to just keep learning more and more and being like Oh I get that or if you do start a project you'll be really tempted whenever you get stuck to just be like ah I got stuck there but I kind of know what I'm doing no don't do that slowly go through the project make sure you know how to do everything because you want a strong Foundation that's going to make everything so much easier when you go forward okay here's a little project to get you started okay assume you're a marketing data scientist and you want to code a function that takes in a data frame and the function should check whether the data frame has a sales column and a website traffic column and the job of the function is that it should be able to divide those two in order to get the conversion rate and then also say whether that rate is trending up or down and that should be in a separate column okay simple project you can do that one to start off with if you get stuck and you've actually been trying try use chat GPT to see if it can help you out and double check whether What It produced is correct by putting in an actual data frame and seeing if it works step 3 maths so now we have the building blocks of python the next step is to look at the mathematical element because once again it's such a broad field maybe your job will need maths maybe we'll need a little bit less maths but being at least okay at mats is very important I have this video over here which goes in depth telling you exactly what mathematical topics you need so check that one out after this so I won't go too much in detail and if you're partner can be like Oh I'm bad at math I'm bad at maths maths is a skill anything that is a skill can be learned and improved okay I don't speak any Italian but if I press practice two hours a day for the next year in a year don't you think I'll be at least okay at Italian it's the same thing for maths except you're much better off because even if you're bad at maths you have a foundation in it from high school and even just life in general okay so don't panic about the maths okay so now that you know these basics of python as well as mathematics just do another simple project more simple projects building up maybe code a calculator python does the calculations for you so just make sure it's a nice input interface and maybe do something like coding a function that manually differentiates a polynomial which is again pretty easy you can do that just to get started off with baby steps and we'll keep building up and up on top of that okay so now we have the basics nailed on so now I want you to do more in-depth projects so again assume that you're a marketing data scientist what I want you to do is go on to kaggle kaggle is a website you're going to have to familiarize yourself with basically it gives you free data so just go on to kaggle type in like marketing data set or something like that and literally just explore see imagine yourself as a data scientist for a marketing company and ask yourself what could the manager ask me to Output from this data and just pretend like they've given you that assignment and create a Project based around that if that feels like too much to be able to dictate to yourself what to do again go on YouTube and just look up a long form tutorial where somebody goes through a marketing data set there's a couple of really good ones um I'll probably link them on screen or in the description just do those step by step you start off by following others and eventually you'll be able to do it yourself other things that are really important just learn how to read documentation for libraries because a lot of times for future projects the more complicated stuff you have to look up libraries which are a little bit less common so you have to be able to read the documentation and interpret exactly what they want that's not a big step it won't take you that long from this point it's hard to direct you exactly what you should do now because you have the basics all you have to do is explore your interests okay so if you want to be a tech data scientist go on kaggle look for Tech related projects or brainstorm stuff that you'd like to know for yourself if you want to be a healthcare data scientist again kaggle Healthcare data set and create a project around that okay and all these projects we're adding to a GitHub so take a little bit of time to learn how to upload your repos to GitHub not too complicated that's something I'm also solidifying in myself in 2023 but honestly in a day probably less than that you should be pretty good with just uploading projects onto GitHub because especially as a self-taught data scientist you're gonna have to have a strong portfolio so after doing two maybe three more larger projects now move on to understanding apis so apis is basically like a fancy Library it just lets you access data usually from online sites like Twitter or YouTube or most websites have apis so even like genius have their own API and I'll be doing a project around that coming soon on the channel and just keep exploring see what data that API can bring you do a project around that and now you're getting familiar with using apis okay after that point I would advise going down the machine learning route at least knowing the basics of how to do some machine learning algorithms how it basically Works what it is and that's such a broad subfield or actual field of data science so you can explore that when you come to it and honestly with just that python and a base level of maths you can start applying at least for entry-level jobs and internships when you're starting off you might not get the ideal job that you want but just get that first job once you have that first job you're on the ladder it's easier to go to the next one and to the next one especially as a software data scientist just make sure that your portfolio looks good and by looks good I don't mean you have 20 random projects have like maybe two or three really good in-depth projects for your portfolio you know the ones we'll work you on before if you can Target them towards the exact sector that you want to apply to that will help a lot and have that on your GitHub and if you have personal website like I do or something like that that will really help to show off what you can do now one thing I will mention and this point is often overlooked by data scientists is learning SQL is really useful not all data scientist roles require SQL but just knowing it and how it functions it's super good to be able to put on your CV because a lot of the times you might get unstructured data or exports of data from different sources and you have to join them you could technically do most of that in Python but honestly just being able to do it in a different language is really useful and honestly I find it easier SQL is a lot easier to learn than python so it shouldn't take you too long so it'll be really useful to put that on your CV all that's left now is to fix up your CV and I'm actually in the process of reworking mine just in case I have to apply for a job you know always going to stay sharp and I will be talking about CV stuff on the channel soon but basically just put your most relevant experience near the top show off your portfolios and really Target it towards the job that you're applying for so that's basically what I think you should do if you want to learn data science in 2023 but there's no one-size-fits-all solution so tweak these things if you feel like learning one aspect first before that one or after that one will be better for you then do it that way okay all that matters is that eventually get to the point where you feel like you can do some data science work if you want more detail on exactly what matters and what coding to do these videos are linked now but besides that I'm data Nash a data scientist on his journey from being a newbie to one day being an elite data scientist so if you want to join that journey and learn from my mistakes see why I succeed at feel free to hit subscribe peace

Channel: Data Nash

Views: 91,960

Keywords: data science, data analytics, data science job, data engineering, tina huang, study md, ali abdaal, ken jee, how i would relearn data science in 2023, how to become a data scientist fast, 2023 data science roadmap, how to learn data science, How to Become a Data Analyst in 2023?, sundas khalid, data analyst road map 2023

Id: Z79AqDouS-Y

Length: 12min 24sec (744 seconds)

Published: Sat Jan 07 2023