How I Would Learn Data Science in 2023? (If I could start over)

and don't know where to start well this video 
is perfect for you because I am going to talk   about how I would learn data science in 2023 if 
I were to start all over again I did a version   of this video in 2022 and I like to refresh 
it every year with my updated thoughts based   on the industry Trends in the data science job 
family so in this video I'm going to summarize   six steps that I would take to learn data science 
we would be talking about what research you need   to do and then we'll jump into what skill set 
you need to learn followed by what do you need   to do after you're done with all of that as we 
go through the steps I suggest you take notes   I'm going to link down a notion template below 
which you can use when researching different   data science roles different data science job 
family and for creating your study plan number   one thing that I would do to become a data 
scientist in 2023 is I would research the   role when I started to learn data science I 
heard that data science is the sexiest job   Emily thanks to the Harvard Business review and 
I jumped in I wish I wish I had done the research   try to understand the data science job family a 
little bit more to understand what encompasses   the job family and what are the different options 
in there before I jump into one in 2023 there are   many roles within the data scientist jaw family 
and I'm going to list a few of them there is data   Engineers machine learning engineer data scientist 
product data scientist data analyst just to name a   few so if you are just starting out and don't know 
anything about these job families I would suggest   that you read up on each of these role and try to 
understand what they do before jumping into any of   the roles that are in the data science job family 
because this step will give you actually a really   really good idea what the expectations of the 
role are and what you will enjoy doing so let's   say you have decided that you want to become 
a data scientist generalist you can think of   it as like full stack engineer who does back-end 
friend that the middle and everything in between   you can think of it as like a full stack data 
scientist that does pretty much everything they   know statistics machine learning they know coding 
they can take the business problem and apply data   science to solve those problems and add value to 
the business so let's say you figured out that   data scientist role is what you want to do you're 
not done researching yet what I would suggest you   to do is go read the job descriptions for the 
data scientists role at different companies   that you are interested in as I mentioned in 
some of my previous videos data scientist is   a slightly difficult for Deaf Family to be in 
because the definition for a data scientist is   not well defined from company to company so this 
is why you need to put a lot of attention on this   step because your study plan your roadmap is going 
to look different depending on what companies and   what specific roles that you're trying to Target 
let's say Amazon is the company that you want to   Target and you want to become a data scientist 
there you will go to the job description for the   data scientist open roles at Amazon and you will 
try to understand what are the skill sets required   second you will go to LinkedIn look at people who 
currently work as a data scientist or have worked   as a data scientist at Amazon and try to look 
at those people's educational background what   did they study then the type of projects that 
they've worked as a data scientists basically   we're actually doing this research and we're 
basically actually doing data science here   statistics we basically ran collecting samples of 
people and trying to understand what are the what   are their educational history what is the type 
of project that they do this is also a good time   to look at if they're like specific certificates 
that they have these people have specific degree   programs specific boot camps that these people 
took that basically helped them so this will get   you actually a lot of information so let's 
say you've done your research where you're   ready for the third step the third thing I 
would do is I would learn the fundamentals   of data science you probably were not expecting 
to hear that because a lot of other advice that   you have heard is basically is telling you to 
start coding I strongly believe that in order   to be a data scientist a good data scientist you 
actually need to have a solid Theory knowledge of   statistics and machine learning before you get 
to the coding part and there's a reason why the   coding languages is a way to apply data science 
they are not the data science itself you can ask   any data scientist who is currently working in the 
industry or is going to school you can ask them   like what is data science they will tell you data 
science is pretty much statistics domain knowledge   and machine learning knowledge so that's why it's 
important for you to build those fundamentals   before you start getting into coding now here one 
thing I would say that if you hate coding then   I would not pursue a data science because data 
science does require coding and if you have never   coded in your life in that case I would suggest 
you to try out coding first before you jump into   the theory but theory is so so important and it 
basically builds the foundation for data science   so I would suggest you to learn statistics and 
learn the machine learning fundamentals before   you jump into coding and we're doing this step 
but we're not going to go too deep we're going   to go at the high level you're going to try to 
understand what statistics actually is you're   going to try to understand what machine learning 
actually is so before you jump into the next part   let's say if somebody comes and asks what is the 
difference between linear and logistic regression   you should be able to explain it that's the 
level of knowledge that I suggest you to have   before you jump into the coding part the coding 
part would actually when you apply the knowledge   you will actually get to learn statistics and 
machine learning much more but for the initial   learning period I would suggest to stick to the 
fundamentals learn statistics mathematics and   machine learning so these are the three areas 
that I would suggest to build your fundamentals   on so here is how I would approach it for math I 
would be very comfortable with linear algebra for   statistics I would get pretty comfortable 
with probability distributions hypothesis   testing Bayesian versus frequencies so those are 
basically the basics in statistics the third I   would suggest you to go into machine learning 
and machine learning try to understand what are   the types of machine learning there is supervised 
unsupervised reinforcement and then within each   try to understand what is a regression what is 
a logistic regression what is classification   what is decision tree we're not going too deep 
but we're understanding it enough that we we   can explain theoretically what they are because 
we're gonna go back to it again now a word from   our sponsors simply learn if you're trying to 
start your career in data science and looking   for a structured program that simply learns data 
science boot camp might be a good fit for you the   program is developed in partnership with Caltech 
University which is ranked number nine in the US   and in my opinion gives that credibility it's 
a six month cohort based boot camp but 25 plus   Hands-On projects going through the curriculum 
I really like that it starts with building   foundation in statistics and machine learning and 
then jumps into coding and teaches python SQL and   R with hands-on experience in three different 
domains also I really like that it focuses on   interview prep which is much needed because we all 
know data science interviews are not easy in this   program you will learn from global data science 
faculty who have combined 40 years of experience   the cohort is starting soon and has limited 
seats I'm linking The Bootcamp below check it   out it might be a good fit for you now back to the 
video so let's say you build your Theory you have   a solid foundation in statistics machine learning 
now it's time for you to go to the next step which   is learn to code there are many languages that 
you can learn for data science for the simplest   it is take I would suggest to start with SQL 
and python I personally started with sqlnr r   has a very steep learning curve and it's not as 
intuitive as python but R has a lot of ready to go   statistics methods available to you that you can 
use for for doing like a very quick analysis but   art cannot be productionized by whereas python 
can live in a production environment I would   suggest you to start with python and SQL python 
is super intuitive and as an industry I have seen   that python has actually been taking off because 
it's easy to understand by everybody involved in   the project including software engineers and 
data scientists for learning SQL and python   SQL you can you can go with any Learning Resource 
for learning SQL it's pretty straightforward you   just need to understand how to join different data 
sets in our right left self joins which is like a   weak point for me how to do SUB queries how to do 
window function for python now there are a lot of   courses that are out there for a learning python 
here I would like you to pay special attention   to focus on learning python that is specific 
to data analysis for example there are several   libraries that are focused on data science and 
data analysis and machine learning such as pandas   numpy scikit-learn matplotlib just be mindful when 
you're picking your Learning Resource for python   make sure that it has a focus of data science 
because there's a ton of learning material out   there that teaches to software engineers and 
non-software Engineers I would suggest you to   focus on python that is more targeted toward 
data science at this point you have a really   good understanding of fundamentals of data science 
which is statistic machine learning math you have   coding knowledge which is SQL and python now 
is actually it's time to apply the skills that   you have learned so far and turn it into a project 
there is a possibility that while you are learning   all those things learning by python learning SQL 
through whatever Learning Resource that you use   you probably already did the projects so you have 
like some hands-on experience but this step is   specifically to build your portfolio and to get 
you more Hands-On knowledge on how to do things   start building your project portfolio and remember 
the first step that we did or the second step that   we did where we were looking at different people 
on LinkedIn who are working at your target company   in that role if you have already written down 
what type of projects that they do in their role   as a data scientist this will give you actually a 
very good idea of what type of projects to Target   what kind of focus areas that you need to have 
in your project portfolio if you're looking for   data there are actually a lot of free resources 
available to you where which has tons of data   and problem sets that you can use to build your 
product portfolio listing a few including kaggle   Google data search US Census Bureau and you don't 
have to be limited to the data that is available   in this platform you can actually make your 
own data for example you can look at your your   purchase history on your credit card download 
that data turn this into a data science problem   and do a project project on it identifying your 
purchase trends for example How likely are you   to buy a coffee if it's raining what I'm trying 
to say here is like you can look at different   data set make your own problem and try to build 
projects around it build at least five projects   get your data from kaggle Google data search 
or make up your own data Target a domain that   you are interested in but also a domain where 
you want to get into for example if you want to   be a data scientist in e-commerce then you would 
pick a data data set that is related to that and   solve a problem that is related to e-commerce I 
would also suggest if you have the option is to   build the online project portfolio link it on 
your LinkedIn link it on your resume and build   a GitHub portfolio this is optional I personally 
didn't do a GitHub portfolio but if I were to do   it again and I don't have any experience I would 
build a GitHub portfolio so recruiters can look   at it and they have additional information on my 
skill set that they probably don't have on other   candidates who don't have GitHub portfolio the 
reason you're learning trying to become a data   scientist is to get a job if you are doing it just 
because you're curious like that's great but most   people who are trying to become a data scientist 
they want to get a job so the sixth step that I   would recommend is to prepare for interviews and 
the reason I say this is because a lot of people   discount how much work interviewing is knowing 
the skill versus doing it in an interview setting   where you are under a pressure environment and 
you have to answer in a time constraint manner I would start practicing on a platform like lead 
code or start a scratch I'm going to link it below   and that is for SQL and Python and then for your 
fundamentals in your theoretical knowledge I would   start mock interviewing and start practicing 
with a friend have them ask you questions so   that way you are ready for the interview itself 
I've created a detailed video on how I prepare   for interviews and what is my process you can go 
and watch it here I'm going to link it somewhere   here it goes in a lot more detail that I'm going 
in this video so hopefully by the end of this   process you are able to go into an interview and 
perform to get a job offer now that being said I   do want to mention that over the last few years 
given that there's so much interest in the job   family it has become more and more competitive so 
don't be discouraged if you don't get your job on   the first try so this is the process that I would 
use if I were to learn data science all over again   do any of these steps resonated with you surprised 
you let me know in comments with that thank you   so much for watching this video and I will see 
you in a different one have a beautiful day bye

Leave a Reply

Your email address will not be published. Required fields are marked *