Session 1- Introduction to Data Analytics

Fundamentals
so this is data analytics fundamentals and we are starting today so this is going
to be the first session and my name is Ragu uh I will I will talk about myself a
little bit don't worry some of you might know me and I may be new to some of
you right um so before we even start U this course like I said this something
very special it's not like a regular course that we do and I know that the
logistics for the course comes up later uh but what I want all of you to do is
quickly grab a pen and paper pen and paper or pen and book.because you will
be doing something in this course I mean like you may not be sitting idle you
may have to write something do something so I'll give you probably a couple of
minutes for that but please please grab a pen and a paper or a book so you can
write something right uh so this is something else we have included certain
activities and we we really really would love all of you to participate in those
activities and for those activities you will need a pen and a paper right so so
usually we talk about the logistics later I mean after couple of slides uh but
I thought I will give you some time to grab this so okay so uh let me talk a little
bit about myself um you may probably figure out that I look a bit older than
from my picture so there are certain reasons for that probably I'll talk about it
later but you know this is my real picture you know what you see here is not
real me it's me but uh I I don't know right so uh to talk to talk a little a little bit
about myself um uh I have around 16 years of uh industry experience and
primarily what I do is I work with data so a lot of people ask me like hey what
do you do like what is your job so um over the period of time there were like
different roles different positions but you know after certain time you really
understand what you do so now I usually tell people that okay I work with data
that's it I just work with data right data is my bread and butter what I do with
data is a different question I mean there are a lot of tools and technologies that
we use but end of the day what I try to do is get the data make some sense
from the data and help people to use that data that's basically what I do uh but
if you really want a technical introduction about about myself um for the last
10 to 12 years uh I have been working in the area of big data analytics so big
data is this area where you handle huge amount of data and then analyze it
and then make predictions and you know insights from the data so my primary
expertise is actually big data analytics and there are lot of tools and
technologies that we use in that and I was fortunate enough to work with a lot
of mnc's uh including Cisco I have worked with Microsoft uh if you talk about
Flipkart uh if you talk about Uber I was a consultant for them so I was fortunate
enough to work with a large variety of companies from the industry as a
consultant uh even in the banking sector I have wor so uh and and that's about
me I think we should speak less about me me and more about the course right
so and if if you really uh talk about the technical uh expertise of me like I said

00:04:08 the primary expertise is actually Big Data uh on top of that I also
work on cloud so basically AWS uh Microsoft Azure Google Cloud I work on all
these platforms and probably a little bit of uh machine learning and AI you
know I I I seriously think about this right you cannot live without ai ai is
everywhere so I also work a little bit on the machine learning and artificial
intelligence site all those things has one thing at the center which is data right
so that's what I simply said I just work with data that's more easy to

00:04:50 convince people right anyway that's my introduction not

something very important and this is our Learning Journey for today right right
the to or the table of contents uh before we jump into this to I have one more
thing to discuss about this course because this course says data analytics
fundamentals so when the first time when we were discussing about this
scores I was like what you want to do a fundamental scores on data analytics
who will do that and then I realized nobody has done this before I mean it's a

00:05:27 surprise right so when you hear the term data analytics what is
that comes to your mind right a lot of tools and Technologies and cloud and
this and that and all these things right or probably Microsoft Excel you know a
lot of tools basically so the interesting fact is that there are lot of courses in the
market where you actually teach a tool for example if you want to master Excel
there are millions of courses to to learn Excel yeah I mean you can find it
anywhere you want or you want to master

00:06:05 Tableau powerbi Hadoop databases you know there are like
millions of courses out there so what I what we figured out is that a lot of
courses are there which teach you the tool but there are zero courses which
teach you what you should be looking in this tools for example example I may
learn Excel right uh I will learn Excel today probably I'll take two months to
learn Excel I know all the shortcuts I know macros I can literally create
everything in Excel but end of the day why am I doing this I

00:06:48 mean if I get a job and if somebody gives me some data what will I
do with it I know all the charts all the formula everything but is there some sort
of of like a ground for all this that's what we are doing here so it's very
important to set the expectation for this course because this course is very
unique it's nothing that we have done before I mean it's like very fresh very
new and if you look in the industry I have not seen this course anywhere you
can probably check it out because this course is

00:07:25 talking about what do you want to do with data it's not talking
about a tool for example we are not doing Hands-On with Excel you know
calculating top 10 Revenue you know calculating the profit no no no no no stop
stop stop we are really asking the question why are you doing this if I give you a
tool if I give you some data how do you make sense from the data what are the
ground rules to approach this problem that's what we are addressing in data
analytics fundamentals so like I said it is a very

00:07:59 very unique course and um I'm quite sure that it will give you a
perspective a fresh perspective to look at the data right and then once you
understand how to make sense from the data you can use any tool it doesn't
matter just to give you an idea uh when I started my career I was using
Microsoft Excel okay to to to do the basic data analytics Microsoft Excel is what
probably it'll cost you I don't know th000 rupees per month right today I mean I
don't know somewhere around th000 rupees after five years or six years I

00:08:40 was using Apache Hadoop which cost $1 million I mean look at the
difference right thousand and 1 million but the same kind of analytics I can do
using both of course the size of the data may vary but I can use Excel and Hado
to draw the same conclusion from the data but to do that I should know what
I'm looking for so this is very interesting it's like a detective story you know you
I don't know if you love this but I really love Sherlock Holmes and Watson you
know Arthur K and Doyle and there is

00:09:16 a series about them right and um BBC has done a Sherlock series
so this is more or less like that okay I'll give you some clues some evidences
some data what is your inference yeah so let's get on with it so that's the idea
why we are doing this course and this is session one and I think uh I did my
introduction that's fine and in this uh particular uh session this is our road map
of the content right uh so we start with um data analytics introduction its
importance and the origin right then we

00:09:55 talk about the role of a data analyst we'll get into that and then we
have some very interesting real life case studies for example one of them is
related to IPL and one of them is related to the uh most rated TV show in the
history of mankind I mean I don't know if you watch TV shows or series which
is the highest rated series that you know yeah so a small example on that then
we speak about types of data analytics uh where there is descriptive diagnostic
diagnostic predictive and prescript analysis so we'll get into

00:10:40 these terms so right now you may be thinking hey this guy just
started the course and he is throwing a lot of jargons at me I don't even
understand what he's saying don't worry this is just the rot map that we have
all right so these are the points to remember I think some of them I already U
explained uh first of all take notes if you want and use a pen and paper I want
all of you to have a pen pencil paper whatever you want just to uh do some
activities so we want this session to be interactive so make sure to answer any

00:11:14 questions asked by the mentor oh I'm going to ask questions not
like that and I just want like the session to be interactive now usually we start
with a slide but I'm not going to do that right so in every class it's like the first
slide what is this what is that I'm not going to do that I'm going to give you
probably four minutes to do an activity I mean you have to do it then I will help
you out so what is this activity all about uh this activity is planning the perfect
weekend so what I want you
00:11:53 to do is take your pen and paper and I want you to plan a perfect
weekend weekend let's take Sunday so today is what Thursday right okay so
this Sunday how do you want to spend your time what are the things that you
are planning to do and how effective can you do them what about the money
right I mean it's not like I want to see BJ Khalifa this weekend I can definitely
see BJ Khalifa but that'll be like too costly have to fly to Dubai right uh probably
I can do that but the money is there but but

00:12:30 instead of that probably I'll watch a movie then go out with my
friends I don't know right so just be honest with yourself uh you don't have to
show me the answer like okay this is what I'm picking my girlfriend and going
here you don't have to do that but just spend uh three to four minutes and
write down how a perfect Sunday will be for you and this is important as an
activity so I'm going to give you like four minutes to do it starting now uh pick
up your pen and paper and try to do it if you're really admin like okay

00:13:05 I'm not going to use a pen and paper uh then you can plan this in
your mind but I really love if you use a pen and paper yeah so let's see how
many of you can plan a perfect weekend yeah let's see see if I ask you for a
perfect weekend you may just say that okay I'll go to a movie or something like
that but let's break down this process right and just be with me so so the first
step that I'm going to do is Define the question right so Define in the sense like
when I say a perfect weekend what does it mean to you is it like I'll go

00:13:43 for a movie not everybody likes a movie is it like I go out with my
friends or is it like I sit and watch probably binge watch an entire series at
home I don't know so the first thing is that we Define what we want so let's say
we are a movie bus and we plan to watch a movie okay so that's one thing that
I definitely want to do and the second thing is that okay after the movie uh I
want to take a short break maybe go home and then in the evening probably I
want to hang out with my friends so this is my primary
00:14:19 idea but at the same time what I also do is I consider other options
so the Second Step that we normally do is collect the data so if I want to watch
a movie probably it's a nice idea to watch Mission Impossible right so that's in
theaters now so I start collecting the data okay what are the movies and where
are they playing how much is the ticket and if I cannot make to a movie what
are the other options I have so probably I will rate them you know watching
Mission Impossible in IMAX will be 10 for me because I don't die hard

00:14:59 Tom crwi is fan right uh maybe something else it may be at maybe
having breakfast or roaming around the city visiting a garden so I I pick up all of
them and I label them and I say that okay this is what I really want to do but
these are the other option so Gathering data is really important about what
you want to do the third step that I might probably consider is clean the data
uh when I say clean the data I may have a lot of unwanted information right for
example I collect I go to book my show or some place and look at all the

00:15:41 you know um theaters where Mission Impossible is there and

maybe there are like 100 theaters I I just want to get rid of all this clutter I
concentrate on probably two three theaters which are my nearby same goes
with visiting a garden or whatever you want so step number three is that you
clean your data or whatever you have and step number four is analyze it so this
is where uh I look at my list and I really want to make a decision for example
um I may want to watch Mission Impossible but then I realize like if I

00:16:19 go to IMAX it's probably too costly I can go to a regular Dolby

Atmos theater that may be a little bit less and anyway Mission Impossible is not
short on IMAX it doesn't make any difference right so I may weigh all my
options analyze it and then I may decide that okay this is where I want to go so
that is Step number four step number five is visualize the results now this can
be done with the help of a tool for a complex problem or this can be done in
your mind if it is something small what you basically do do

00:16:57 is that you take a pen off uh pen and paper and then you write
down everything and then you plan your day for example you say at 9:00 I
wake up and 10:00 I'm out 11:00 I'm in so you basically visualize what you want
to do now you may be thinking hey for a simple week and why are you writing
all this down I mean what difference does it make you see the difference
previously most of you told me that I'll just go watch a movie or something like
that and that was very vague now you break down that same process into
different steps so you

00:17:33 get more options to compare and also now you are more certain
about what you really want to do so why I'm picking up this example is to
demonstrate that data analytics is not some Hightech you know super
computer thing it is something that we apply to our daily lives so this is a
classic examp example for data analytics to plan a weekend you may be
thinking planning a weekend has nothing to do with data analytics but if you
really apply these rules and write it down I mean think about it right you may

00:18:14 just want to go out for a weekend but what if you have 10 friends
or 20 friends everybody will have a different opinion right so rather than just
fighting with them if you just lay out like this then a decision can be made very
easy right for example somebody may suggest I'll go to this movie this location
you can actually compare and say that hey this is going to be less cheaper or
you can say that this to this time we are free so we can probably use it for
something else right so data analytics basically is a part of our

00:18:50 Lives it's like gravity not exactly like gravity but see gravity is
something which is acting everywhere but you really don't believe it right I
mean you believe it but uh if I drop a glass it hits the floor because of gravity
but I don't think like okay gravity is pulling the glass right but gravity is
everywhere same like that data analytics is part of everyday life even people
who have never seen a computer apply data analytics on daily lives so this is
the most important point I see in a lot of

00:19:27 interviews I mean these days so we I tend to be in the panel of lot

of Camp companies for interviews and in a lot of interviews the interviewer will
be asking these kind of questions given a problem how do you break it down so
this was just a sample dose for you I mean if you were sitting in an interview
and the interviewer asked how do you define a perfect weekend what will be
your answer I'm quite sure you all of you answered really well but what the
interviewer might be really looking for is a stepbystep drill down approach that

00:20:05 we created so that's why this is important so kudos to all of you I

mean there is no right or wrong in this I just wanted to do a small activity
before we jump into the slides and this is the introduction to data analytics so if
you look at a definition of data analytics it says uh data analytics is is the
science of analyzing row data to make conclusions about that information it
involves using various tools and methods to identify patterns Trends and
relationships within the data that can help organizations

00:20:45 make informed decisions and improve their business operations

now honestly speaking this looks like a Wikipedia definition to me right so we
have to have some definition right uh so this definitely looks like a definition
but you know I don't want you to like okay by heart a definition or mug uper
definition so let's understand what this definition is all about so the first thing
that it talks about is it is a science of analyzing row data to make conclusions
about that information now I'll give you a simple

00:21:27 example let's say you have a piggy bank and you're collecting
money not now but let's say you are a kid and you're collecting money right
now I have a uh eight-year-old daughter niney old actually and she has a piggy
bank she's she collects everything in that piggy bank coins and notes and
everything and when I look at her sometimes she breaks the piggy bank opens
the pigy bank and says that okay you know what I have more coins than the
currency not and sometimes she say that you know what last month was the
month when I got more

00:22:04 amount of money and sometimes she says okay Dad you're not
giving any money for last two months that's perfectly data analytics so it's
basically deducing something from row data so the piggy bank is an example
for row data it has just coins and currency notes but from that you can identify
Trends and patterns and everything right for example my daughter identified
that last two months I was not giving any pocket money from the piggy bank so
she was actually making sense from how much she is saving and all so that's
that that's

00:22:40 but but basically what you mean by identifying patterns Trends
and relationships right and why do we really do this so that we can improve the
business now end of the day the goal of every organization is is to improve the
business right uh so if you take this into a different example uh one of the
example that comes to my mind is McDonald's you see McDonald's is a very
large food chain right so they have like thousands and thousands of uh Outlet
right especially in the US they have thousands of outlets so what McDonald's

00:23:21 does is that they collect the sales data from all these outlets right
so they collect the data first that's the row data for example today we sold
thousand Burgers something like that then what they do they also collect the
weather data what was the temperature what was the humidity then they do a
comparison to identify trends for example I don't know how many of you know
this I really love the ice cream in McDonald's oh my God I mean that that small
ice cream what you call cone right I mean it's

00:23:57 like very cheap cheap and very nice so McDonald's can actually
figure out which are the days of the month or which are the months in which
soft drinks and ice cream sales are high right or which are the months in which
Burger sales are high so this will help them to adjust their operations for
example if McDonald's already know that in the month of March people are
going to buy more ice creams and soft Rings they can prepare their employees
for that they can tell their employees that okay you know what you are going to
get a lot of

00:24:35 orders for them so basically data analytics is a science where you
convert row data into some sort of relations and Trends so that organizations
can make money end of the day they can make money I mean yeah so end of
the day we I mean I always this is a very funny thing actually I mean the end
goal of all this is like companies make money I mean yeah we say operational
expend expenditure has decreased or whatever but they make money yeah but
that's why we are here right I mean otherwise there is no point

00:25:10 anyway so that's data analytics introduction and why is it so

important right so like the last part of the last slide right okay we collect the
row data we make some inferences but then why are you doing it right so the
first point is datadriven decision making uh like the McDonald's example you
know all your decision can be made using data second is enhancing customer
experience now there is a very interesting example for this uh you know
Walmart right Walmart so I mean I always give this as a very

00:25:51 interesting example because when we think about data we always

think about uh you know numbers and you know Excel sheets and tables and
database and all right but there are other type of data as well so for example in
Walmart what they do is they have these cameras everywhere okay and when
you walk into a Walmart store you will be recorded you will be tracked and why
are they doing it so what Walmart basically does is that they try to analyze
what people are trying to pick up for example let's assume that there is a
Walmart store in

00:26:36 Albuquerque does that sound a does that ring a bell Albuquerque
we are coming to Albuquerque in a moment okay so let's say there is a
Walmart store in albuquer and lot of people are visiting the store so probably
most of the people are buying vegetables we don't know right so what
Walmart basically does is that they track these people and they realize how
much time they are taking to reach an aisle for example vegetables may be way
back that means the customer has to work a lot to reach there right on so if
they

00:27:15 analyze this data I mean the video data they can figure out which
are the items most frequently picked up by customers and then they can
rearrange their shelves so that the customer experience is improved so if I'm a
Walmart customer maybe I also buy a lot of vegetables next time when I visit
the vegetables aisle is in front of me so I will think okay nice I mean I don't have
to walk all this way and I don't know if you have visited a Walmart store it's like
really huge you will practically get

00:27:52 lost inside one right so so that's an example for uh customer

experience where they use a combination of video footage and artificial
intelligence to rearrange their shelves uh increasing operational efficiency
again a lot of companies uh what they do is by looking at the sale of the
product uh they can um you know control their logistics for example there is
Pepsi and Coca-Cola right and some people like Pepsi some people like Coca-
Cola what do you like I like Coca-Cola for some reason so Pepsi and Coca-Cola
they're

00:28:31 like huge companies right I think both are under one umbrella I
don't know whatever it is right so if you can figure out what people like in one
particular area then your Logistics can be adjusted for example I know in this
city people are going to consume more Pepsi so there is no point in
unnecessarily shipping a truck of Coca-Cola I mean people love PC anyway right
in that City so that's operational efficiency and predicting Trends and Market
changes so I'll come to this predicting Trends later I mean

00:29:07 prediction is a difficult word to talk about I'll come to that uh

mitigating risks and frauds especially in the banking if you look at the banking
right I don't know if you have ever noticed this right uh you can try this if you
want I mean don't try but if you want you can try right if you have a credit card
for example right and if you swipe your credit card for a bigger amount
immediately a customer care agent will call you have you ever experienced it I
have done that for example let's say I

00:29:43 have a credit card and the limit is 5 lakh rupees I swipe for four
lakh for some reason immediately they will call I mean I have experienced it
many times the second example will be the same credit card if if I use for an
international transaction like Euros or US Dollars immediately they will call why
so that is mitigating risk and fraud all these credit card companies and Banks
constantly do data analytics to figure out fraud so I mean what are the chance
I'll swipe for four lakh rupees very rare right unless I don't know something

00:30:25 I mean very rare right so if there is a transaction of 4 lakh rupees

right there is a chance it is a fraud and to understand that you have to analyze
the previous data data analytics right you have to analyze all the data to come
up with this and personalization and targeted marketing so that's basically like
giving you customized offers and all right say for example you buy stuff from
Amazon or let's take a different example you listen to songs from Spotify a lot
right so you know that if you keep

00:31:02 on listening to Spotify they give you personalized recommendation

right if you're a premium user of Spotify the recommendations are even better
because you're paying money right and they want to retain so how do they do
this how do they give you personalized recommendation that's with the help of
machine learning but end of the day that's part of data analytics so today I
cannot think about any business which does not involve in data analytics every
business does this now there is an interesting slide I mean doesn't mean

00:31:40 like the other slides are not interesting but see um if you think
about data analytics so I asked this question to many people like okay when do
you think all this started like the data analytics right uh so can you take a guess
like when did data analytics actually started I'll give you like 30 seconds you can
say like 1980 1940 1900 I don't know any any year when did data analytics
actually start what do you think it started in the Paleolithic Era that's like 1
million year before can you imagine you thought

00:32:22 it started with computers right you're absolutely wrong even I

thought like that so one of the early signs of data analytics or analytics so we
call it is the ishango bonds in the Paleolithic Era you can actually Google this
right so this was like a earlier form of calculator so and this were used by very
primitive people like imagine 1 million years ago how was people right I mean
like very primitive but even they had analytics in their brains so I think it is
more or less passed by the genes and
00:32:59 you know all I mean it's it's it's it was already there so the first uh
glimpses or the first evidence for data analytics is actually from Paleolithic Era
then if you look at ancient Egypt and ancient suaria that's like 3,000 BC they
were very good at data analytics for example in Egypt they were doing regular
census and statistics yeah even statistics dates back to 3000 BC right and you
know this pyramids in Egypt right and pyramids were built using humans we
believe we really don't know how they were built but it took a

00:33:44 lot of efforts to build uh pyramids so in early days in Egypt they

were using stats and senses to allocate Workforce for building pyramids data
analytics in The Mummy Returns right I mean yeah so Egypt was there and
then suaria right so ancient suaria they had clay tablets as databases right I
mean when I say database what does come to your mind I mean Microsoft seel
and people are like okay I know all this but 3,000 BC people had databases not
like what we see today but of course write scribbles and then

00:34:24 tables but ancient Sumerians were very good at tables and
correlations uh but now we have to come to the present day right in the 19th
century one of the early uh examples for data analytics was actually Henry Ford
the person behind Ford motors and he was using data analytics to optimize the
assembly line you know car assembly line and in 1940s government started
using computers for predictive analysis is in many government areas and in
1960s uh we started using something called decision support systems or DSS

00:35:08 so basically DSS is like large computers but they are more into
predicting like what is going to happen right so you don't have to byard this it's
like it's not like somebody going to ask you like what was the thing they were
using in 1960 this is just like like a timeline now if you look at today what's
happening so in the late 20th century Big Data came that's my area obviously
so big data like I said is huge amount of data I mean if I ask you what is the size
of data you normally say gigabytes you will say terabytes but
00:35:46 then if you go beyond there is like petabyte exabyte so th000
terabyte is petabyte thousand petabyte is exabyte so when you have row data
in these volumes how do you make sense how do you do analytics that's where
Big Data comes into picture and uh after that in the early 21st century uh it
became very critical right and then the emergence of AI in data analytics so the
latest Trend I should say is AI or artificial intelligence and you know you will
hear a lot of terms related to data and unfortunately I may not be able to

00:36:31 clarify all of them because they are beyond the scope of this
course for example there is data fabric data mesh data Lake data warehouse so
many things right but all of them represent taming the data storing the data
and making sense in one or the other way that's basically what it is so we are in
2023 and now now data is a complete different ball game data Remains the
Same but now we have the power of computers and it is like increasingly ever
now we have the power of computers and we have the help

00:37:13 of machine learning and AI to do data analytics I'll speak briefly

about this uh later like machine learning and AI uh but but yeah so that's
something that I want to talk about what does a data analyst do so remember
we took that example of uh what you call the planning a perfect weekend right
so those are the step typically uh involved in any data analytics area right for
example broadly if you look we first Define what we want to do the problem
which is very important then we collect the data in

00:37:53 various forms I'll talk about it uh then we clean the dat data we
remove unwanted data so on and so forth and then we actually analyze it and
then we visualize uh but if you look at a data analyst let's say you are a data
analyst uh these are some of the thing that he does first of all you are
responsible for data quality and integrity right so data quality is a very
important thing for example you may be collecting data from lot of sources
how do you know if the data is correct what if you are

00:38:30 getting the wrong data how do you understand it right so having
the quality of data is very important a simple example for this is uh I have a lot
of experience working with sensors you know the temperature sensor uh you
know atmospheric pressure sensor so a lot of sensors are there so what
happens when these sensors become faulty so when you collect data from
sensors you know what happens uh they will give you the reading so let's say
you are collecting temperature so they will give you okay today the
temperature

00:39:02 was 30 degree 29 degree whatever but one problem is sometimes

these sensors become faulty like we had an experience of collecting uh sensor
data the temperature data basically and normally when you collect these uh
temperature data you actually get the reading right for example 30° 25°
whatever it is right but sometimes what happens the sensor goes bad right so
when the sensor goes bad it gives you a default reading so what you get is
basically like maybe 15° celius right so you know that uh it's

00:39:44 not correct the actual temperature is not 15° but that's the default
value you are getting so you have to ensure that okay the quality is there in the
data right you you need to know which which data to take and which data to
avoid basically that's one of the rule then analyze and interpret complex data
to provide insights obviously that is the job create reports and visualizations to
communicate findings right so all data analysts are familiar with visualizations
in one or the other way um you may probably use Excel or

00:40:27 PowerPoint or probably a complex tool like powerbi or Tableau to

create visualizations and this usually is charts right you may be familiar with bar
charts and pie charts and whatnot right so that's one thing that uh data
analysts are really good at work with various stakeholders to understand uh
data needs use statistical tools and techniques to analyze the data so the data
analyst also works with uh stakeholders what do you mean by that an example
will be let's say you are working with bhi AEL you know airel right

00:41:13 and one of the challenge will be something what we call customer
churn so probably you figured out that last month a lot of customers you lost
so why are you losing customers probably Reliance Geo came up with a new
plan you don't know right so you have to analyze the data and you have to
come up with a reason and you have to present it to the stakeholders when I
say stakeholders this can be company owners this can be the board members
because they will ask you right okay why are we losing customers right

00:41:55 so a data analyst also closely works with stakeholders right and
this is a typical process followed by uh data analyst I'll give you an example to
uh reiterate this idea this is like just a slide so this is the same steps that we
basically followed in the planning the perfect weekend activity so we first
Define the question like in the case of bhi airel the question will be why are we
losing customers that's the question you need to you need to U answer the
question why are we losing customers Second Step will be collecting the data

00:42:38 so first you will be collecting data from your customers the
feedback from your customers Market data your competitors like Gio or any
other vendor how are their plans how many customers they got last month
that's collecting the data then you clean the data you take only what you need
and now you analyze the data you may use any tool to analyze the data and
you create a report in the form of visualization where you say that okay you
know what last month Gio introduced plan number one two three and we
don't

00:43:18 have a competitive plan so we so these plans offer I don't know

games for free yeah and for this reason uh a lot of youngsters migrated
because they love games maybe I'm just giving an example right and then you
also make recommendations you tell that okay how do you mitigate this
situation what is the solution right the solution is either reducing the plan cost
or coming up with a new plan for only teenagers something like that right so
even though data analysis sounds very simple the role is not very

00:44:00 simple yeah so when we say data analyst people think okay so he
just analyzed the data but a data analyst role is very important in an
organization and he touches each and every aspect of the company I'm not
talking about a beginner probably like if you're hired as a beginner probably
you won't do that but once you become exper experienced right you will have
your say in everything so let's take a simple example to understand this right in
fact let's take uh two examples uh one is a cricket team example right so let's
say

00:44:44 you are Amani maybe not Amani right somebody you know you
are somebody who is really interested to pick up an IPL team and we are not
talking about the usual teams like uh you know Chennai Super Kings or
Mumbai Indians because those teams are real but let's say a fictitious team
right so what is that you have to do you have to pick up the perfect cricket
team right so how do you pick up the perfect cricket team right so you can see
here there are different steps involed in this for example there

00:45:22 is context description key element what's involved and real life
impact so you can think about this like uh how do I say this like a framework or
a set of rules to help you pick up the right team so what is the context here the
context here is that you are an investor and you want to invest your money
wisely and the description of the problem is you want to select the best cricket
team now what are the key elements that will impact you decision making on
data the player data understanding of constraint

00:46:09 so if you look at this from our earlier example the first step will be
data collection right so we collect the data about the players their background
their previous performances we collect data about pictures different St
stadiums competitors all this we collect so that is the data collection part right
and then what we do we clean the data and then we analyze the data that all
comes in your key elements here and then there is what's involved right
making decisions on which player to include in

00:46:47 the team based on available data and constraints for example you
may really like a player to be included but due to let's say the money constraint
you probably don't want him right I mean uh you can't just have everybody
right on the other hand you may probably find another player which is not who
is not so uh uh you know good at batting or bowling right now but after
carefully analyzing you saw that he's really good he had a great carrier and with
a little bit of help and coaching you can make

00:47:24 him play so these kind of real life examples are there and what
could be the best example apart from Ms dhon I mean Mahendra Singh donon
you love him you hate him you cannot ignore him that's the truth right I mean
I'm a huge fan of Ms D there are a lot of haters obviously but whether you love
or hate you can't ignore him and donon is a classic example for data analytics
now you may be wondering how donon is an example for data analytics I'm not
saying that uh after a match donon sits with the laptops and runs some

00:48:08 algorithms or anything he does everything naturally I mean he's a

natural player and a natural data analyst right now let's look at this from a data
analytics point of view right what is the context here real life example of
strategic decision making in cricket so what is g ton so good at strategic
decision making right so he always refers to previous matches and
performances and and then makes his decision right so do you remember this
uh donon had this decision to use joinder Sharma in 2007 T20 World Cup now

00:48:56 Jinder Sharma was a relatively unknown player but donon had
done a lot of analytics I mean like Mental Math probably and and came up with
this idea to actually use jindra Sharma and it played really well now the
description for our problem will be how D's strategies reflect principles of data
analytics the key elements are like I said understanding past data right so so
donon looks at a player and then looks at his past data and then decides
whether he should be included or not included or etc etc the

00:49:39 second thing that gor is very good at is predicting future outcomes
right so he knows how to shuffle batting order and and you will be amazed to
see this I mean I have seen this hundreds of times right I'm not saying like
every time D wins but depending on the situation he can predict the batting
order sometimes he's number third sometime he's number fifth even others so
predicting the outcome uh in D's way is probably reshuffling the batting order
third is making data informed decisions right so he actually there was
00:50:21 an example right uh he had to promote himself in the up in the
order in 2011 World Cup final and we won right do you remember the 2011 uh
World Cup finals and in that he used data and he promoted himself in the
batting order because he knew that he had to bat early and that contributed
and we won and continual monitoring and adjustments so donon constantly
adjust his strategy I is based on you know the team the conditions everything
right so like I said uh this is more or less like solving a murder mystery in case of
a

00:51:06 Sherlock Holmes mystery right you get the data you look at the
past you try to look at the future then you catch the thief in a mystery but here
you don't catch Thief but you make the right uh you know predictions right
then what is involved right so historical data is involved predicting outcomes is
involved and real life impact is successful Cricket strategies leading to victories
so this is what we call a framework now a framework is something that a lot of
people doesn't know what it is right you

00:51:45 hear the term a lot but people really doesn't know what is a
framework a framework is a set of rules to approach something in data
analytics the framework that we use is what you see here so there will be a
context first that is like what you want to do how do you want to do there is a
structure then there is a description like in our D example the description is like
how donon strategized everything then what are the key elements inv worlded
you know collecting the data looking at the past making the decisions

00:52:24 Etc then what's involved right what are the actual steps involved
and what is the real life impact now I always say this data so a lot of people tell
me that okay you know what Ragu like data is gold mind data is nothing I I
don't believe data is gold mind that's incorrect actually I know that lot of you
may not agree with me data is not gold mine if you make sense from the data
then it is gold mine so there are lot of organizations who collect the data and
they sit on top of the data and they don't know what to do data is there
00:53:05 but they don't know what to do that doesn't make any sense right
so that's where this framework comes into picture if you get some data you set
all these context description key elements what's invol real life impact and then
you have a stepbystep process to approach the data right now I wanted to give
you one more example right I mean I don't know if this example this is a
general example for uh maybe data analytics right how many of you are fans of
Breaking Bad how many of you have seen breaking bad now why am I talking
about

00:53:47 breaking bad because data analytics can be applied to Breaking

Bad okay what is the context so let's look at the different uh uh steps here right
the context here is you want to make a series and the best series that's
breaking bad right and what's the description of what we want to do the
description has to do with budget you see when breaking bat started it has five
seasons if you not seen it it's okay right season 1 was like a I don't know like
they were just trying I mean know nobody was sure it'll be a hit or

00:54:27 miss so season 1 was okay in season two what happened they
went way out of the budget yeah so there was a budget for a season right so
let's say for example 1 million was the budget they cross 1.5 million in season
two and then season 3 came and the showrunner Vince Gilligan in an interview
he said that their biggest challenge was was to keep uh everything within the
budget so what is your description of the problem is keeping the budget right
now how do you keep the budget so what are the key elements here right so
first what they

00:55:11 did they analyzed the previous data season 2 Data so they realized
that in season two they had 13 episodes in season 3 also they have 13 episodes
and they did some analys is and understood okay these are the areas where we
spend a lot of money and these are the episodes where we had lot of location
so money was gone etc etc then they cleaned all those data right and now they
needed a solution so in season three of Breaking Bad again you cannot cut
down the number of episodes if you look at it you have
00:55:52 13 episodes in season 3 but after do carefully doing the analytics
and comparing the budget with the previous season they came up with what
we call a botal episode you know what is a botal episode okay I'll tell you there
is an episode in season 3 which is called the fly you remember that episode
right fly a fly will enter into the secret lab of hurg and Jesse and all they do is
try to catch the fly yeah there there's nothing else in that episode so that's
called uh bottle episode a bottle episode is

00:56:32 something where you minimize the cost the location everything if
you look at the fly episode there are only two people Brian Cranston and Aon
pole and the entire shoot is in the lab but they did it so well that that episode
became really good nobody understood but they minimized the budget the
budget came down a lot only because of that particular uh episode so you look
at the approach right first you look at the historic data then you understand the
problem then you realize you cannot cut down the

00:57:08 episodes then the solution from the analyst will be okay let's do a
bottal episode so one episode we are going to shoot in one location just two
characters cost drastically comes down that's breaking bad for you so my point
is it is applicable everywhere if you look at it right and people often mistake
these terms there is like analytics analysis and analyst I mean yeah so analytics
analysis and analyst so analytics is the systematic computation analysis of data
or statistics right so that's what we were

00:57:50 talking about right we are talking about data analytics so the first
thing is analytics analysis is actually different it's different actually so it is the
process of examining data in order to understand and interpret it right so this is
like breaking down a very complex topic into smaller parts right an example for
analysis will be like if you look at these labs right some of the labs they will be
conducting a lot of experiments now these experiments as a whole may not
make any sense so they break it down into smaller part so that

00:58:39 everybody understand it that's analysis an analyst is a person who

actually does analysis right so this can be a financial an analyst or a business
analyst or anybody like that so this is just like good to know information
nobody is going to ask you like okay tell me the three difference between
analyst analytics and Analysis but just these terms are a bit confusing that's
why and this is probably the uh last slide we have I know that people might be
thinking what is this and this is really important this is called the

00:59:16 types of data analysis so there are normally when you talk about
the types of data analysis we normally say there are four descriptive diagnostic
predictive and prescriptive I know that sounds like very weird but let me give
you an example descriptive is like what happened in the past so a retail owner
might look at last quarter Revenue to understand whether it is good or bad or
something that's descriptive descriptive analytics B basically means what
happened in the past anything diagnostic means investigating

01:00:03 why it happened right so for example like in my bharti airel

example uh last month we lost lot of customers that's descriptive why it
happened because Gio introduced a new plan that's diagnostic third is
predictive predict means what is going to happen in future so if I'm working for
bharati airel I will say that okay if you don't do anything you are going to lose
more customers next month that's predictive prescriptive analysis means your
recommendation so what do you want to do so basically you can say that okay

01:00:45 introduce a new plan or do whatever right so I have a very uh I

mean I have an example for this from from my experience okay so this is like I
was working with the flip cart okay and when I was working with flip cart we
did this right so uh one of the problem we had was about orders getting
cancelled so a lot of customers were cancelling the orders why so the
descriptive analytics was We did an analysis and we figured out that okay in the
last three months a lot of cancellations happened that's

01:01:28 descriptive what happened diagnostic was like why the customers
are cancelling orders then we figured out that in the last three month we
changed the career so I think I was working with flip cart in 2018 they had tied
up with one of the career company so that change happened in the last three
month so when the career company changed what happened happen the
number of days for the delivery increased for many customers Yeah so basically
I order something and I have to wait like seven days so I cancel because I'm
getting

01:02:07 better deal on Amazon Predictive Analytics was that if you keep on
doing this then you are going to lose money prescriptive was the solution to
this right so we had a meeting with the corer company and they had set up last
mile delivery centers I mean in some of the outskirts they didn't have delivery
so what they did they set up small warehouses and all so that customers can
easily get the packages so you see you can break down almost any analytics
into these four category and that basically is the end

01:02:47 of our uh today's class but uh let me also quickly do a recap so we

started with uh data analytics remember and there is the introduction
importance and then origin we spoke about all that and then we discussed the
role of a data analyst right and then we looked at uh some of the uh you know
uh real life applications like IPL D example Breaking Bad Etc and then different
types of data analytics right uh then descriptive diagnostic predictive
prescriptive that's what we [Music] discussed