Learn data science for beginners (How to learn data science for free)?

I get these questions from the people

saying that I'm a mechanical engineer or

become graduate can I learn data science

the answer is yes I can say for sure you

can learn data science and the reason

I'm saying this is because I know few

people in my life who have become a

successful data scientist or a software

engineer and they come from a totally

non-technical background and these

people are not genius they're average

people so if they can do it you can also

do it in this video I'm going to show

you how you can learn data science

effectively using a step-by-step

approach this video has two section in

the first one I will go through all the

topics that you need to learn are for

data science and the second section will

show you the resources that you can use

or to learn these topics now it all of

course starts with a programming

language data science is a process of

drawing insights from data and you need

programming language to operate on this

data Python and R are the two most

popular programming languages used by

data scientist if you ask me for my

preference I would always go with Python

because once you learn Python you can do

a full stack application development now

if you try to master Python it might

take a long time so you want to learn

only few topics in Python and then you

want to move on you don't want to be

stuck with Python for one entire year so

these few topics are variables and

numbers the basic data types such as

strings lists dictionaries control flow

structures such as F and for loops

functions and some basic understanding

of Python modules as well as reading and

writing files once you know all these

topics you are ready to move on to the

next big topic which is numpy and pandas

numpy and pandas allows you to do data

cleaning and data exploration data

scientists pan

majority of their time in data cleaning

and data cleaning is very essential

because the data that comes in real life

it's very very messy and you want to

clean that data drop unnecessary

features handle any values and so on

and for doing that numpy and pandas are

very very useful in numpy and pandas

also you want to only focus on few

topics such as data frame and series

basics data frame is the main object in

pandas that allows you to represent the

tabular data you want to know how to

create a data frame from various files

such as excel csv or even from SQL table

then you want to know how to handle any

values in data frame and panel data

frame has nice api's which allows you to

handle any values then you want to learn

how to merge and conquer the data frames

how to do group by and how to do

different I SQL type of operations on

these data frames now of course when you

are coding or you need code editor to

write the code and these code editors

are also known as IDE integrated

development environments I prefer

PyCharm or Visual Studio code as my code

editor so you can pick one of these two

and then along with this you also need

Jupiter notebook Jupiter notebook is

little different than IDE

it allows you to write code as well as

do interactive data visualization so I

find myself using Jupiter notebook all

the time and in addition to not book of

one of the code editors either

pycharm or vs col is required data

visualization is a technique used by

data scientists for doing it exploration

and for that you can pick either

matplotlib or c bond these are the two

Python libraries that allows you to do

data visualization in these libraries

also you need to know

the basic chart types such as line chart

bar chart pie chart you need to also

know histograms for plotting the

frequency distribution knowing axis

labels leigh-jensen grids allows you to

plot your chart in a beautiful way and

then of course cattle plot is important

in plotting your data into two

dimensional space so once you know these

basic type of charts you are ready to

move on to the next topic which is SQL

as a data scientist you are always

dealing with data you are retrieving

data from some data tables and for which

you need to use structured query

language and you need to have a good

understanding of relational database

let's say there is already a big

database within your company and as a

data scientist now you want to get that

data to do your data analysis and data

might be scattered into multiple tables

now you need to know how to perform join

between the tables how to run select

queries where queries you know order by

and group by clause so just this basic

understanding of SQL will help you a lot

with our data collection as well as data

analysis you of course need to have good

understanding of math and statistics as

a leader scientist when you are dealing

with huge volume of data you want to use

some statistical concepts such as mean

median normal distribution standard

deviation for removing outliers etcetera

and some basics about probability and

descriptive and inferential statistics

all these topics are going to help you

in your data analysis process when it

comes to math of course linear algebra

or differential calculus knowing all

these topics are very helpful now if you

got let's say 2 out of 50 in your maths

class doing a school list don't worry

too much about it

Matt is something that you can learn

gradually so if math was a weakest point

don't think that I cannot become data

scientist that is not true you can

improve your math gradually by referring

to the math from a good resources for

example one of the good resources three

blue one be due to channel that person

explains math in such a beautiful poetic

way that math doesn't sound that complex

okay so if you are patient enough and if

you take step by step approach then you

can conquer math as well

then comes machine learning of course as

a data scientist you will be building

machine learning model if you're using

Python SK learn is the library that

everyone uses for machine learning in SK

lon you need to know how to build a

regression and classification model okay

these two are supervised learning

techniques other than that you can learn

unsupervised learning techniques such as

k-means in order to convert your text

data into numbers because machine

learning models understand only numbers

you need to know some encoding

techniques such as one-hot encoding or

label encoding train displayed a k-fold

cross-validation grid such CV all these

are essential in evaluating model

performance and deciding the best model

and doing hyper parameter tuning so you

need to have a good understanding of all

these concepts as well then comes deep

learning for deep learning the two most

popular libraries or frameworks are

tensorflow which is from google and pi

touch which is from Facebook there is

theano as well from Microsoft so you

need to know one of these frameworks and

just get your understanding clear on

what is neural network what are like

hidden layers and output layers what is

activation functions what are different

types of neural networks CNN

and so on now one last very important

tool I want to mention is Microsoft

Excel whether people like it or not

Excel is still being used in data

analysis when your data size is small

and if you don't want to write code for

doing data analysis Excel comes really

handy in Excel you can set different

data filters you can apply formulas you

can even plot different charts use

vlookup and pivot tables it has a rich

set of functionality which allows you to

do data analysis in a quick way in this

diagram here I have outlined all the

tools that we discussed I'm going to

provide a link of this entire

presentation in video description below

so you can refer to all these topics or

later on as well so don't forget to

check the video description other than

these tools when it comes to Enterprise

2 there are for example there are things

like tableau and power bi other people

use for our data visualization there is

Hadoop and spark that data scientists

use for storing big data and for

distributed computing Amazon sage maker

and Google cloud computer lao's you to

do a cloud base a machine learning

computing so all these are enterprise

level tools that if you know it's of

course well and good but when you're

starting your data science journey you

don't need to know them you can just

start with the initial set of tools that

I mentioned and then when you have time

you can learn these enterprise tools as

well

now whatever tools and technologies I

mentioned an effective way to learn

these tools is a project-based learning

approach what I mean by that is you

start out with a project and then in

order to finish that project whatever

technology and tools are required you

learn them step by step this way you

have a concrete goal in mind and while

you're learning all these topics you're

also slowly making a progress in your

project for example you decide to build

a website for whom price prediction now

you take this project as your end goal

and in order to finish this project

whatever you need to learn Python numpy

machine learning you just learn them

step by step and you make gradual

progress in that project and in the end

after two or three months you would have

finished the project as well as learn

all these different topics require for

data science for this particular project

which is home by PI's prediction I have

a complete tutorial playlist so you

should probably go through that where I

have shown how to build that website

where there is like HTML CSS JavaScript

UI there is Python flask server and

there is a machine learning aspect

machine learning model building aspect

to it there is data cleaning there are

just this project just covers pretty

much all the topics which I require in

data science so I highly recommend that

you go through this tutorial list now

let's talk about from where you can

learn all these different topics

starting with Python I have a nice

Python tutorial playlist in this

playlist you can follow up to 16

tutorials I would say and that will get

you started this is the tutorial design

for an absolute beginner so you should

be able to get good understanding of

Python basics by following this tutorial

then comes pandas for pandas also I have

a very nice tutorial playlist where I

use of feather data for doing pandas

manipulation here in this list also you

can follow till 9 tutorial ok so just

follow this 9 videos and after that you

are good to go

then comes Jupiter notebook for that I

have simple 3 tutorial playlist which

will give you idea on what is Jupiter

notebook how you can use it how you can

install it so if you're not familiar

with this notebook concept then this

will be the best place

to get started then comes a matte blot

lip for matplotlib also I have a 7 video

playlist and these are like very sort

and simple videos which will give you an

understanding of how to do plotting and

data visualization in matplotlib for SQL

I like this guy could have anchored he

has a nice tutorial playlist so here you

can follow till I would say joins so up

to 12 so just followed the up to 12

tutorials and that will give you a good

understanding of what is relational

database joins and various SQL concepts

ok now all these tutorial playlists that

I am mentioning I am going to provide

link of them in the video description

below so don't forget to check video

description it has tons of useful

information when it comes to statistics

I recommend this think stats book it's a

free PDF book available on internet so

you can just download it and you can get

your statistics concepts clear one other

place to check out is can academies

statistics and probability course so

this has like nice video or tutorials so

you can just follow that for machine

learning I have another playlist which

will give you a basic understanding of

supervised on supervised our learning as

well as one heart hot encoding k-means

crits or CV all those cool concepts so

in this playlist I recommend that you

follow all the tutorials because those

are all very useful it has three deep

learning tutorials as well and then

there is like a data science projects

where I show how to build a website for

home buyer price prediction are step by

step for deep learning there is another

Coursera course which is available for

free for Python there is a book called

automate the boring stuff with Python

this is also available freely online for

digital copy is available for free for

Python data science that is this book

called Python data science handbook that

is available for free on internet so

check this out it has covered all the

like numpy and pandas and data cleaning

and all those aspects it has covered

pretty much in depth you can see that it

has machine learning also there is an

EDX course for Python for data science

this is another great course you can

check it out if you want to on a

certificate then you have to pay the

fees otherwise just to watch videos and

going through the course there is no

fees there are a couple of other EDX

courses this one is data science

essentials from Microsoft then there is

data science foundations and then there

is learning from data so I hope that

gives you some direction on how to learn

data science if you have your own ideas

or any free online resources that you

are aware about which can help other

people then post them in a video command

below