How I Would Learn Data Science in 2022 (If I Could Start Over) part 2

hi friends welcome back to another video in 
today's video we're going to talk about how   i would learn data science if i could start all 
over again this is part two of the first video and   if you haven't seen the first video where i 
actually shared my path i highly suggest that you   go ahead and watch that video because this video 
is going to build on top of it one of the piece   of feedback that you all gave me in that video is 
that i should be highlighting the path with some   resources for you to learn from so in this video 
we're going to do exactly that i'm going to share   how i would learn data science with the resources 
that i would use for learning data science and   thanks to coursera for sponsoring today's 
video and making it possible for me to share   these tips with all of you today so since this 
video is sponsored by corsair the courses that   i will be mentioning today is specifically from 
corsair i understand there are many resources out   there that are available to you in this video 
i'm specifically going to focus on coursera   courses because personally i'm a huge fan of the 
offering by coursera so let's get into the video   so i'm gonna go step by step in terms of how i 
would learn data science what is the first thing   i would do what is the second thing i would do and 
then in each of those steps i will have resources   linked again like if you haven't watched the 
previous video i would highly recommend that   you go watch that um if you haven't watched it 
i will cover i will be covering some of those   topics that i shared the the steps i shared at 
a high level so the first thing is let's say   you want to get into data science the first thing 
that you should be doing is finding your niche as   i mentioned in my previous videos data science 
is a huge job family so i highly highly highly   recommend you to narrow down your focus and study 
all the jaw families that are in data science so   you can narrow down your focus and figure out what 
is it exactly that you want to do in data science   so some of example of data science job families 
include data engineering data science data   scientist generalists machine learning scientists 
applied scientists research scientists and so many   more so the field is huge and there are several 
roles within it so the first thing you should be   doing is narrowing down your focus and figuring 
out where do you want to pursue a career because   that will make your study plan much easier once 
you know what you want to do you'll be able to   define your path much more easily create a study 
plan around it and follow through so let's say   you figured out what you want to do uh let's say 
you're picked data scientist generalist as your   niche and you want to continue making a steady 
plan for that so the first thing i would recommend   let's say a data scientist generalist is your 
target role one of the piece of advice that a   lot of people give is to start learning to code is 
the first thing you do i disagree with that advice   and here's why so i'm gonna i'm gonna say this 
coding languages and programming is a method is   a tool to apply data science it's not the data 
science in itself and so for that reason what   i recommend to start with is learning the basics 
which includes the fundamentals of statistics and   machine learning so you're getting your foundation 
right so think of it like you're building a house   you want to set the foundation first and what 
is the foundation of data science or a data   scientist role the foundation and the core work 
and understanding of a data scientist role is   application of advanced statistics and machine 
learning so when you start with the foundation   you learn the fundamentals start with statistics 
start with math start with machine learning   that way you have your foundation set so whatever 
you do next you're building on top of it so   now what i promised that i will be sharing 
resources so here are some of the resources that   i would personally use for learning statistics and 
machine learning and getting a fundamentals right   one of the course that i'm actually personally 
taking right now as a refresher is introduction to   statistics by stanford this is an amazing course 
and it covers the basics it covers the descriptive   statistics probability data sampling distribute 
data distribution regression inference and more   and as i said i'm personally taking this course 
and i'm actually really enjoying it and i really   like the simplified explanation in this course 
and i truly believe that once you when you   start learning these basic statistics knowledge 
and then go into advanced statistics you're   actually getting a better understanding of 
what data science actually is the second is   i would focus on machine learning so a lot of the 
concepts that you would learn in statistics would   also apply to machine learning so once you start 
doing machine learning you will actually realize   that a large chunk of machine learning is driven 
from statistics and for machine learning one   of my favorite data science instructor is andrew 
angie who actually recently launched a new machine   learning specialization course on coursera on june 
16th in partnership with deep learning ai and this   is actually the upgraded version of the previous 
machine learning course that was available it   covers various topics in supervision learning as 
well as unsupervised learning including multiple   linear regression logistic regression neural 
networks such as deep learning decision trees   unsupervised learning clustering dimensionality 
as well as some of the best practices used in the   silicon valley and another bonus of this one is 
that instead of using the matte matlab which is   what the previous course ml course was using this 
one is actually has assignments graded assignments   and lectures that teach python so these are 
a lot of topics and this course is actually   11 weeks and it's pretty rigorous so it's possible 
that you'll start taking this course and you're   like oh my gosh what did i get myself into it is 
pretty intense so i would make sure that when you   go into it you go in with your full attention 
and you're able to grasp those topics and if   you get stuck like there's a community and that's 
another thing that i love about cursora because   there's like a community of people who have taken 
those courses and then there's a discussion forum   where you can ask questions you have access to tas 
so you can ask those questions you can you have   access to other students who are taking the course 
so you can do a lot of like peer learning from   each other so hands down a great machine learning 
course definitely something to look into and   let's say if these two are not enough and you're 
still trying to build your foundational skills   um your basics of data science um and let's 
say if you want to get a bigger picture   of what data science is and how it's applied in 
the industry ibm actually has an awesome course on   coursera as well which is actually a certificate 
so ibm data science professional certificate that   course goes into what is data science what is the 
application what are the different data science   methodologies so this course actually covers the 
whole data science the data science job family   and gives you a very nice overview it gives you 
an overview of what data science job family is   and then you get a very good understanding of 
how data science work in the bigger picture and   the last part of the course also taps into a bit 
of python so the three courses that i mentioned   these are not hands-on coding courses these are 
courses that are more theoretical and that's   the part that i want to emphasize on building your 
theoretical knowledge is very very important 90 of   these courses are more on theories and the theory 
can be dry i i know like coding is a lot of fun   and you like you get to do hands-on work so this 
can be a little bit dry so make sure that you're   prepared to be slightly bored and if you're 
not getting bored which is actually perfect so   it it might it's possible that it it might be a 
little bit too much at once so pace yourself so   you're able to grasp the knowledge and retain it 
to sign up for any of these courses i mentioned   click the link in the bio and i'm going to 
list all the courses that i mentioned today   in the description box so make sure that you 
click there and you sign up for those courses   on that note i'm going to tell you a little bit 
more about coursera coursera is an educational   platform with hundreds of courses and certificate 
from top universities and companies designed for   individuals as well as organizations you get to 
learn from the comfort of your home from top tier   faculty which is one of my favorite parts 
and in fact my top favorite data analytics   course by google is actually available through 
corsair corsair offers both free and paid options   and one of my favorite parts about coursera is 
that it offers financial aid so if somebody cannot   afford to take that course a paid course you can 
actually apply for financial aid which means that   education on coursera is accessible to everybody 
and let's say if you're planning to take multiple   courses on carcera you can also look into crusader 
plus which gives you access to over 7000 courses   at a monthly flat fee which is a pretty sweet 
deal so talking about retaining the knowledge   let's say you build your theories you build 
your fundamentals you build your theoretical   knowledge you have very good grasp of what machine 
learning is what statistics is you understand the   basics you have a good foundation of theoretical 
knowledge and if given a data science problem   you're able to solve it using theory so now the 
next part and this is the fun part and i know like   a lot of us like i personally love enjoy coding so 
this is the next part so let's say once you build   your fundamentals the next part that you'll jump 
into is learning to code and this is where you   actually apply the knowledge that you have learned 
and get hands-on experience i have done a detailed   video in terms of the languages that you should 
be learning as a data scientist i'm gonna link it   here and i'm gonna link it somewhere here you can 
go ahead and watch it but specifically i want to   focus on three languages that any data scientist 
generalists should be very very familiar with   number one is sql second is python third is r so 
you can pick a combination of sql python or sql r   i'm going to share courses on all three of 
these and it's actually up to you to decide   which combination you want to pick do you want 
to pick sql and python or do you want to pick sql   nr i have my own opinion in terms of what language 
i would recommend i personally would go to python   and i have very legit reason of why i would choose 
that i've done a detailed video on that as well   so you can go ahead and watch it um but i'm gonna 
mention courses for all those three languages all   right so as i said i'm gonna mention courses on 
all three languages so let's start with sql which   is my favorite i think a lot of people there is a 
misconception that data scientists do not need to   know sql i actually completely disagree i think 
you need to have a very very good understanding   of sql because you will very likely be using sql 
to access your data do the joins and to make sure   like you have your basic data ready before you 
actually do any advanced statistics or machine   learning on it so sql is a must-have so for 
learning sql i really like the course learn   sql basics for data science fundamentals by uc 
database and what what i love about it again like   there are so many courses on sql but what i love 
about these courses that i'm specifically going   to mention today for sql python r is that these 
are designed for data science in mind so these   are for people who want to get into data science 
these are for people who are learning data science   so this course specifically is targeted toward 
analyzing data with in mind that this is for   your learning sequel for learning data science 
it actually has a hands-on project there are four   courses in the course and if you finish all those 
four courses you can actually earn a certificate   in sql so which is great so the next is python so 
python again it's a very popular language there   are so many courses that are focused on software 
engineering as well as data science so you have   to be mindful which course you pick one of the 
courses that i've recommended to a lot of people   for learning python is python for everybody's 
specialization by university of michigan   this is created for data science in mind and there 
are over 1 million people who have actually taken   this course so it's a very very popular option and 
people keep taking it because it does a wonderful   job teaching python um it's again hands-on course 
which involves a lot of quizzes and uh projects   and a lot of it covers python data structures 
how to use python to access web data which is   super cool because you can do project on any if 
you learn how to access web data you can actually   do project on any of your favorite websites how 
to use python for when working with databases   and then the final course is the capstone project 
which combines everything that you've learned in   the course again this is just one of the 
course in python that i'm mentioning here   obviously like one course is not gonna 
be enough i would highly recommend that   you possibly supplement this with additional 
learning material as well as when you learn   something make sure that you're doing additional 
projects on top of it to retain that information   so for our programming again there are so many 
options the good thing about our programming is   that most of the courses and learning material 
that are out there they're targeting toward   data science and data analysis which you can go 
wrong so there is a specific course on cursera   that teaches r in a very short it's a very short 
course it's three courses it's the the course is   called data analysis with our specialization and 
the reason i specifically love that that it's   only three courses so it doesn't go too deep so 
you can let's say if you're doing sql and python   combination but you still want to get a flavor 
of r you can still take this course and kind   of like get an idea although i do not recommend 
that you try to learn too much at once because   then you'll just be overwhelmed but let's say if 
you're still curious and you want to understand   how r works and like how to um how to basically do 
the data analysis using r this course specifically   does a good job because it covers introduction 
to probability and data using r so it basically   covers the basics it also talks about inferential 
statistics which i think is is amazing because i   do a lot of experimentation work and inferential 
statistics is very very important so this this is   a good way to also see an application using r 
the last one is linear regression and modeling   again like this is a short course so this is not 
going to cover everything so i recommend that   first of all you build solid foundation on these 
three concepts that i share in these three courses   that are in this course and let's say if you 
finish all these three courses and you want to   get a certificate for it that's also an option 
in coursera let's say if you're taking multiple   courses on coursera and it starts to get expensive 
you can do that you can continue that so it would   be an economically better option to sign up for 
coursera plus which is like a monthly membership   but you get access to over 7000 courses with one 
flat fee which is a pretty good deal it includes   many of the courses that i mentioned today sign up 
for corsair plus visit the link in my description   all right so i think we got the meaty stuff out 
of the way you know the fundamentals you know   your coding you're able to in in step one what you 
learned you able to take your theory you're able   to take your coding knowledge and you're able to 
like do hands-on projects and solve data science   problems so that's where the third part comes in 
taking these courses is not going to be enough   you actually have to do hands-on project to build 
your portfolio and retain that knowledge i'm sure   like you have heard of that term um tutorial trap 
which basically you take course after course after   course and you think like you've learned but 
you haven't when what you actually go to do your   project you're not able to do it that's called 
tutorial track because when you take a course it   you think that you have you have gotten it down 
when it comes time to actually do it you're not   able to do it because there's so many nuances that 
you can run into when you are actually doing the   project so that's why i want to emphasize building 
your portfolio because there are two benefits to   it one it's gonna help you retain the knowledge 
and two it's gonna help you build the portfolio   with additional projects that you would do that 
you can highlight on your resume and it's gonna   help you with your job search and hopefully help 
you land a job and let's say if you have done   the courses that i mentioned previously there are 
capstone projects next to each course so possibly   by the end of taking all those courses you already 
have a project portfolio but i still recommend   that you do more projects on your own and add 
to your portfolio beyond the capstone projects   and another thing talking about capstone projects 
the reason i love courses that have capstone   projects because it the capstone project it gives 
you an idea how a data science project looks like   so you can take inspiration from it and actually 
pick one of your own projects and frame it into   a problem similar to the capstone project that 
you took there are two ways to build your project   portfolio one is that you can go to so many 
existing websites that are already out there where   there they have given you the data set and they 
have defined the data science problem and um you   can go to those websites pick up one of any of the 
data science problems the data set that is related   to it and then start working on your own and solve 
that problem one of the favorite websites that i   like to mention here is kaggle it has thousands of 
free data sets that you can use for building your   data science portfolio the second option might be 
more interesting to you if you're curious and have   a creative side so with this one you can actually 
pick your own problem and you can choose to solve   it for example i love doing content creation 
so one of the projects that i would like to   do which i don't know if it's possible i 
need to think more about it is predicting   if a youtube video will go vital what are the 
elements of a youtube video going viral and then   doing some sort of predictive analytics on 
top of it to figure out if a video given   elements xxx will go viral and if you if 
there is an interest for you let's say cars   video games tweeting whatever it is like you can 
pick an area that interests you and you can um   define you can define your problem and then you 
can work on it using the skills that you have   learned and solve it and that the bonus with 
that is that when you put it on your resume   and you go to an interview you'll actually have 
something interesting to talk about something that   you are actually excited about so something to 
think about all right so we have covered so far   picking your niche learning the fundamentals with 
statistics machine learning and math learning to   code building portfolio and then we're at the last 
part we're not done yet so for those of you i mean   i personally if i'm learning data science my end 
goal is that i want to land a job so i'm sure a   lot of you are learning data science your end 
goal is to find a job right so i don't want you   to stop after building the portfolio there is more 
that you need to do in order to meet your end goal   if finding a job is what it is which i hope it 
is um and i understand if you're just learning   data science if you're curious that's fine too 
you can stop at the previous step all right so   the reason i wanted to dedicate a section to 
job hunting is because interview prep learning   knowing data science being able to do it is 
not enough to land a job and that is because   interview in itself is a monster it requires 
skills for you to be able to answer questions   in a timely fashion so most of those interviews 
are like 45 minutes or one hour and you're given   multiple problems in that time frame and you 
need to be able to solve and allah it will be   theoretical it will be also hands-on coding so it 
requires skills and practice to be able to answer   those questions in a timely fashion under a 
pressure setting so once you're ready to go into   the job market make sure that you're dedicating 
enough time to prepare for the interviews   because it's an investment similar to the previous 
steps i mentioned this is very important because   it's going to help you land the job so i've 
done a few videos on interview framework that i   have i personally use so you're welcome to watch 
those videos if you're looking for more detailed   guide on interview prep it's not specific to data 
science but it's more general you can watch some   of my previous videos where i haven't talked 
i have talked about interview prep and maybe   i'll do more videos on that topic and share some 
resources with you all so stay tuned for that so   this is the honorary mention of one of the courses 
that i was debating if i should mention it or not   i'm gonna mention it anyway because i personally 
love love love this course yes this is google's   data analytics certificate i understand this is 
targeted the name suggests it's targeted toward   more data analytics career letter but i think 
it's super helpful for anybody who is entering   into data science as well and the reason for that 
is let's say if you don't have any data background   if you're just entering new to the data science 
domain this course actually does a very good job   of connecting the dots of why data is important 
and why there are so many jaw families in the data   world that exist today it also does a very 
good job of providing the business context   solving problem and storytelling which i believe 
is super important for a data scientist as well   so although this course is not targeted toward 
data science it does give you a very good   understanding of the fundamentals and the whole 
data world and why we're doing what we're doing   basically so i highly recommend that you check out 
this course i love this course so much i actually   have done a dedicated video on the course 
itself reviewing the course in detail and   giving my thoughts so the the course basically 
has eight courses starts with building foundation   and the course is actually designed in a way 
that is how a data project would be designed   for a data analyst but it gives you still it like 
gives you a very solid idea so it starts with like   foundation of data how to ask the right question 
which again is very relevant to data scientists   as well although the set of questions would be 
different how to prepare your data for exploration   how to process your data from dirty to clean data 
visualization and storytelling so if you're new   to the data domain this course is also actually a 
good start in addition to the other courses that i   mentioned to kind of get a good understanding of 
the data domain itself and why it's important to   ask good questions and why it's important to have 
a have a good storytelling great communication   um and the bonus in this course which i 
haven't seen in a lot of other courses   is that it talks about interview prep they're like 
mock interviews at the end of at the end of the   course so it basically prepares you for more than 
just the core curriculum it gives you like toolkit   to be successful in the job search phase as well 
so which is another reason i love this course   again i wanted to mention this course because this 
is like hands-down one of my favorite course if   somebody is entering in the data domain regardless 
of withdrawal family i think this is a good   start to kind of like build your foundation in the 
data world well i didn't mean for this video to be   this long but i really hope the information that 
i shared in this video is helpful to you if you   have any questions or suggestions for future 
video make sure to give this video a thumbs   up and let me know in comments and again like 
the courses that i mentioned today they're all   linked in the description thank you so much for 
watching i hope you're having a great day bye

Leave a Reply

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