SDS PODCAST EPISODE 87 WITH DEEPAK PRASAD · guys would benefit from Qlik,” and then your goal is...
Transcript of SDS PODCAST EPISODE 87 WITH DEEPAK PRASAD · guys would benefit from Qlik,” and then your goal is...
Kirill: This is episode number 87 with Tableau and Qlik
Intrapreneur Deepak Prasad.
(background music plays)
Welcome to the SuperDataScience podcast. My name is Kirill
Eremenko, data science coach and lifestyle entrepreneur.
And each week we bring you inspiring people and ideas to
help you build your successful career in data science.
Thanks for being here today and now let’s make the complex
simple.
(background music plays)
Welcome to the SuperDataScience podcast, super excited to
have you back here, and today we've got an interesting
guest, Deepak Prasad. Deepak is a Tableau and Qlik expert,
and he works for a company called ASG Group, where he's a
consultant, and he helps other companies to implement
business intelligence in their business. And what business
intelligence basically entails is developing business
intelligence dashboards and tools which can be helpful for
executives, managers, and just generally other people in the
business to see their data and understand how the business
is progressing in various areas, whether it be finance or HR
or perhaps operations, and so on.
So we'll be talking a lot about business operations, and
Deepak will give you some of his insights about Tableau,
Qlik, and Power BI, and we'll touch on a couple of other
tools as well. And of course, he'll give you some tips on his
best practices in business intelligence, so stay tuned for
that. And, finally, Deepak will share parts of his journey and
how he structured his career in this space, and how it
brought him to where he is now. And he'll give you some of
his best tips on how to have a fulfilling career as well.
So I can't wait for you to check this out, and without further
ado, I bring to you Deepak Prasad.
(background music plays)
Hello ladies and gentlemen, welcome back to the
SuperDataScience podcast. Super excited to have you here,
and today we've got an inspiring guest, Deepak Prasad, on
the show. Deepak, welcome. How are you going?
Deepak: Fantastic and really energetic to talk on this podcast.
Kirill: That's good. Where are you calling in from?
Deepak: I am signing in from Sydney.
Kirill: Sydney. Ok, so it was really cool how we were talking before.
You mentioned that you would never live in Melbourne
because what? Why wouldn't you live in Melbourne?
Deepak: Basically, I come from a tropical climate. So anything below
20 is a bit of a pain for me and very tough to accommodate,
so I don't want to take risks in that aspect at least.
Kirill: Yeah, totally get it. Totally get it. And I really want to share
with everybody the chat that we had with Deepak just before
the show started. Deepak, you have a baby boy, right? 2
years old?
Deepak: Yeah, his name is Aryan.
Kirill: What is his name?
Deepak: Aryan – A-R-Y-A-N.
Kirill: Aryan.
Deepak: Yeah, exactly.
Kirill: Wow, that’s so fantastic. Congratulations. It’s always
exciting to hear when people start families. And you were
like — Deepak said to me, “Kirill, there’s one thing that I’d
like to one day find out.” What did you say, Deepak?
Deepak: I really wanted you to get married and I wanted to see you
again after marriage, whether you’re able to tie up with this
podcast and regularly do productive stuff or not. So I just
wanted to put you under an experiment that’s called
marriage.
Kirill: Well, hopefully it won’t be an experiment. It will be, like, one
day for good. But you want to see how my productivity will
drop after marriage, how it will change?
Deepak: No, I really want to see whether it’s going to drop or it’s going
to increase. It’s either way.
Kirill: Yeah, okay. I found that funny. And we both agreed that it’s
a matter of priorities. Once you’re married, there’s just other
things that you care about in life. You know, right now, for
instance, I love travelling. I love doing sports and cool stuff
and just seeing what life has to offer. But once you’re
married, it doesn’t mean that you stop, or you become
boring, or you’re less productive, or whatever. It just means
that there’s other things. Like, you might spend less time on
creating products or courses or something like that, but you
spend more time with your kids and family, and I think
that’s a very, very noble and important and loving and
fulfilling thing to do. How do you feel about that?
Deepak: Very true. Just imagine me after a tiring day at work
returning home, seeing my wife and kid playing and having
fun, and my kid jumps on me and says he missed me and
stuff. That keeps me going. That whole day, whatever
tiredness I had will just vanish as soon as I see my kid
laughing. So things like that will keep you going. Although
you say you cannot do certain things and you cannot plan
and you cannot act as per your plan, but things like
happiness, seeing happiness from your kid and your wife
keeps you going.
Kirill: Yeah, I can totally imagine. I feel the same about my family
and my friends. Yeah, I can totally imagine that. You
definitely need those parts in your life. Otherwise, if you’re
just working all the time, you’re never going to be happy.
You might be successful, but you’re never going to be
fulfilled. I think this is even more important than work.
In fact, I’m actually in the Czech Republic right now, and I
have a friend here, and we just had this conversation today.
Even though he’s not super successful in his career and he
hasn’t made millions of dollars or billions, he’s super happy
because he’s travelled the world, he’s done what he’s
wanted, he’s got a lovely girlfriend, he’s in love and he’s just
living a very fulfilled life. It was very inspiring to see.
Deepak: True, true. You’ll get into the game, man. I wish I could learn
from you.
Kirill: Thanks, man. So, Deepak, tell us a bit about yourself. I’m
looking at your LinkedIn and it says you’re a Tableau and
Qlik intrapreneur. A lot of people have heard the word
‘entrepreneur.’ What does ‘intrapreneur’ mean?
Deepak: Intrapreneur, as the word says, is like promoting some
products and tools across various organizations. So, when it
comes to technology, you will become an expert, and once
you become an expert, you understand its pros and cons.
And once you understand it, you get to expose it to the
people, you get to expose it to the right persons. So that’s my
day in and day out work. I go to people, talk about their
problems and I tell them that this is what will work out after
listening to them. If I straight away go and sell my product
just because I have it, that won’t work.
Kirill: No, it makes total sense. So you’re kind of spreading the
data analytics and data visualization culture within your
organization, is that right?
Deepak: Yeah, exactly. Basically letting the people know the power of
visualization.
Kirill: Okay, that’s really cool. And how is that going? Are you
finding a lot of resistance, or are people generally quite open
to learning about these things that you’re sharing?
Deepak: When we talk about visualization tools, and when we explain
how it works and how it can help, although it’s not the first
thing, people will come and conclude the conversation with
“How much does it cost?” That would be the first question or
the last question. Everyone would finish it. So I always make
sure that the old technology and old processes, what we
listen — and we should not be suggesting some “old process
plus new technology”, as quoted by the HP CEO, I believe.
So it should not be the case that old process plus new
technology is equal to expensive old process.
So I make it as a statement in my everyday work. I should
not suggest something which is just a new technology alone,
I should improvise the process and the behaviour of people
of any organization and then I should show them an insight.
If you just throw a new technology at people, it will not work
out. You should understand their data culture. You should
ask them more questions, and you should ask them what
they’re currently doing and you should really consider them
with a current process and the technology you’re trying to
bring in, is that the value proposition that’s going to be good
or not. That’s the equation you have to arrive.
Kirill: Okay, gotcha. I’m just trying to rethink is. Help me out here.
So, you are a consultant, and you go into other
organizations and then you look at the processes and you
decide that “You guys will benefit from Tableau,” or “You
guys would benefit from Qlik,” and then your goal is for
these organizations to adopt these products. Is that correct?
Deepak: I would say 70% correct. The reason is I wouldn’t just talk
about Qlik and Tableau. Since I’ve been in visualization for
five long years, I know Qlik, Tableau, Power BI and many
other tools. We also evaluate other visualization tools in the
market. For example, I’m keeping an eye on tools like
ThoughtSpot, which is really a disruptive innovation, as I
call it. It is definitely going to turn the industry upside down
one day. So, some tools like this, you have to keep
evaluating and understand the pros and cons.
So when you said Qlik and Tableau, I would say no, those
are not my only verticals. Because I have seen the pinnacle
of Qlik and Tableau, so I call myself a Tableau and Qlik
intrapreneur, but if I feel that Tableau and Qlik will not fit
an organization considering their current process, I would
say that “Qlik and Tableau will not fit. Try something else,”
which means I can suggest these things and these are the
services I can provide you.” So that’s how the dealing should
be.
Kirill: Okay, gotcha. I’m just trying to compare that with what I did
at Deloitte. It’s a bit different because there, we used
Tableau, or some other visualization software at Deloitte,
within the organization, in order to then solve some
challenges or solve a case for our client and then deliver the
results. We didn’t actually go in and say, “You need
Tableau.” We did the work inside and then we just delivered
the results. It might have changed since then, or maybe I
just worked on those types of projects, but that’s my
experience. But for you, it’s a bit different. You not only
deliver the project, you actually want to help the
organization change their processes so that they can further
down the track keep doing this on their own.
Deepak: Yeah, that’s true. The other thing when it comes to business
intelligence, as you know, it’s also called a Decision Support
System, so if something is supporting your decision, it has
to come along with your culture, behaviour and other stuff.
So, for me, data visualization is not just a technology. It
blends people well together, then only you can win. So
anything you produce and no one is there to view your
product, whatever it may be, rocket science for you, but if
the user utilization is 2%, what’s the point in coming up
with a rocket science which no one uses? So when there is a
demand, you should push it and you should create a
demand. The behaviour of the people has to be initiated,
elevated in the right way such as that they adapt to the new
technology and they improvise the process and everything.
Kirill: Okay, sounds like quite an involved process, you would
spend quite a bit of time. On average, how long does it take
for you to complete one of these projects?
Deepak: When it comes to project, I wouldn’t call it as a project. I
would rather put it as a “solution”. So you talk to persons,
you understand what they do day in and day out, and then
you understand what problem they are facing. Once you get
to know their problems, then you get to know what are all
the possible solutions you can come up with. There might be
a scenario where a solution can just be implementing a
product like Qlik or Tableau, or you might be saying that the
way you ingest the data and massage the data itself is
wrong, so we will give you a resilient database and we will
give you a fantastic data mark which consolidates the data
from your whole organization and brings it to a single place
and then we will plug in a visualization tool like Power BI,
Qlik or Tableau, whatever. So it will be an end-to-end
solution. Along with the existing process, we will try this.
Then we will have a discussion with them stating, “Okay,
this is all possible and this is not possible because of a
budget constraint,” or whatever the constraint may be. So
we have to talk like this and we have to do a proper gap
analysis and then come up with one approach which brings
a smile on both our faces.
Kirill: (Laughs) That’s a good way of putting it — one approach. So
you just bring smiles to people’s faces. I think that’s the best
job. It’s like Santa Claus, you’re like Santa Claus of data
science.
Deepak: Yeah, that’s what we are here for. If you just give what they
ask for, smile will not come. If you give them something
more than what they ask for, that’s when the smile comes,
and you see a happy customer, and you’ll also see many
returning customers. When you give something, you should
not just give what they ask. You should give an impact,
that’s what I call it. Impact is what is very important, rather
than just “Give a solution.”
Kirill: That’s true. And what part of your job is education?
Obviously, you need to educate your clients, or your client’s
employees, about how to use this product and how to get the
most benefit out of it and get the correct insights, not the
incorrect ones. So what part would you say is education?
What role does education play in your career?
Deepak: Education is very important when it comes to business
intelligence and data visualization. The reason is, as I
mentioned before, you have come up with some product and
if there is none to use it, no matter what your product does,
it goes for a toss. So, when you do something, people have to
be perfectly educated about what you did and how they can
use it. So when it comes to — I’m a strong believer that
when you design a dashboard, or when you come up with a
product, that product should be following a certain
methodology.
For me, I consider all the users as a king. For example, let’s
talk about a king’s life. If a king wakes up from his bed,
before his feet touch the floor, there will be sandals kept
there. When he walks down to his bathroom, automatically
all the soaps, towels and everything, whatever is going to
work today will be ready and lying there. Just tell me one
thing: Would the king even know who has kept and when
they are kept and why they’re kept. He doesn’t have to worry
about these things. He just has to get into the bathroom,
and he will take his towel.
According to me, the user interface and user experience has
to be like this. For the user, it has to be very natural. You
shouldn’t even realize that these things are all very well-
planned and kept here by someone and that’s when we have
to use it like the soap. It has to be very natural for the
people. That’s why I call it cognitive. So, the natural way of
digesting things. You have to design your products such that
it is very natural for people to use. If you keep your search
bar somewhere in the bottom down, definitely people are not
going to find that at all. The search bar, by default, it has to
be in the right top or somewhere that people focus by
default.
Kirill: Yeah, I totally understand. So where they’re used to seeing
the search bar — on most websites the search bar is at the
top and so on. And at this point we’re slowly venturing into
product design. Before we continue with that, can you
explain a little bit for us? For people who are not really
familiar with business intelligence, what is a product in
business intelligence? What are these things that you deliver
to your end users? What do they look like? How could you
quickly describe them to us so that we can have an image in
our heads what a business intelligence product is?
Deepak: So, as I mentioned earlier in the podcast, business
intelligence according to me is a decision support system.
When you claim that you’re a hero and you don’t have
proper data for it, you are just a person with an opinion. I
strongly believe in that. For your organization, when you get
into a diagnostic mode and see how your business is
performing, and get insights and plan your future, you need
data. So once you have your data, you need a presentation
layer up above what I would call a ‘business intelligence
layer’ where you can ask questions to your data and interact
with your data and come up with your findings to construct
your future.
So that is the layer I’m talking about, that is the layer which
is called business intelligence, which is very important for
each and every organization. So if you just have data lying
around and if your data keeps on growing and if you don’t
ask proper questions in the right time to your data, just
imagine about your organizations. It would be just flat. It
would be just following whatever has been followed, and it
would just keep on doing that. There would not be any
optimizations. There would not be any study of what
happened, what is happening and what is going to happen.
To answer about all these W questions – what, when, where,
why – you need data. So when you have data to ask this
question in a seamless way, you need a business intelligence
layer.
Kirill: Gotcha. So, in a way, a business intelligence product brings
together data of your organization and presents it in an
easy-to-read matter. For instance, in Tableau, that would be
like a Tableau dashboard, in Qlik, it would be Qlik
dashboard or Power BI dashboard… It’s mostly dashboards,
the way I imagine it, and everybody’s seen these. There’s a
pie chart, there’s a bar chart, there’s an area chart at the
bottom. And from that, just by looking at it, you can tell,
“Okay, we’ve sold this many products. Revenue is up, this is
down,” or whatever. Or it could be a segmentation
dashboard for a different department. Is that about correct
in terms of a visual image of what a business intelligence
product looks like?
Deepak: True, very true. When you said, “From various departments
in a single dashboard,” that talks about it. So, when you
have different departments and different data laying around
in various departments and it is not segregated, then you
will not get insight. When all your data is segregated from
various departments and you put it on a dashboard and you
understand the associative relation between each of your
departments, then comes the insight. So that insight to you
can be provided by business intelligence layer is what I’m
saying.
Kirill: It totally makes sense. That’s pretty cool. What we
mentioned, that in different departments you have different
data and different goals. So maybe the finance department
wants to know information on the sales and the revenues
and expenses and so on, whereas the operations department
in a big company that has a call centre, the operations
department wants to have a dashboard on how many calls
are they getting, how many calls did they get this week, how
many calls did they answer, how many calls they didn’t
answer, how many calls were escalated, how many e-mails
they got, they only get information on that, so completely
different purposes. Do you have a favourite? What is your
favourite type of department to work with? Or tell us maybe
about the differences of working with finance, operations or
other departments and the different type of dashboards that
you’ve created throughout your career?
Deepak: For me, HR is really, really interesting data, always, because
I’ve done a lot of things in HR analytics. When it comes to
HR, the way of you seeing the data would be completely
different. What you just described, sales, operations. It’s all
about numbers, right? But when it comes to HR, it’s about
the people. Salary would come from a different department
or whatever department they’re tagged to. And the company
has its own KPI of deriving how people are performing, how
productivity is there.
So, when it comes to KPI in HR – KPI is the key performance
indicator – what I try to say is, when it comes to HR, it is not
about the KPI alone. It’s about people. You have to take
people into the measure. There are many other parameters
that should come into play when it comes to HR department.
That’s the reason I like HR data a lot more. Whatever the
dimensions you see the data, every time you will find some
interesting insight. And those insights differ from various
organizations. Organization A might not be having the same
insight in the same situation as for organization B,
depending upon their HR wireframe and so on. That’s
interesting data, I would say.
Kirill: Okay. That’s really cool. And with your experience in
dashboards, you obviously know some best practices and
how to create these business intelligence products. Can you
share some insights with us, some tips that you can give to
our listeners about what are some tips on creating these
business intelligence products so that they are indeed, as
you said, useful to the people? So maybe how much
information there is, how many maximum charts there
should be on a dashboard, or what colours to use or
anything from your experience that you think is important,
some important guidelines to follow when creating BI
products?
Deepak: Yeah. So, when designing a dashboard, when it comes to
best practices, you have to keep certain things in mind. The
first thing, and very important thing, is increasing the data-
ink ratio. So when it comes to a dashboard, it is very clear
that we are going to put some insights for your data. Other
than data, nothing else should speak in your dashboard.
That’s one primary important thing that I had to highlight.
Let’s say that you come up with a very good insight in a bar
chart and your title is 16 font size. Just tell me whether
people would look for insights in your bar chart or in your
title?
Kirill: In the title because it’s so big.
Deepak: Yeah, definitely. It’s so big. So, you have to be really
concentrating in your dashboard where you want the
people’s focus to be and you have to be carefully drafting it.
And let’s say that you have five or six charts in the same
dashboards and you really feel that you need all the five or
six charts. Situations in this place would definitely ask you
— if I am sitting beside you, I would ask you to revisit again.
If six charts are there, which would be the first one people
would be looking at and whether — when correlating all the
six charts in the same dashboard, you should be making
sure people should not derive some wrong insight that you
don’t want them to. And when you come up with six charts
in the same dashboard, it’s about landscape as well, how
your real estate is going to be maintained and whether
people will be able to focus.
As I told you earlier, the data-ink ratio should be increasing
and the space between each and every object, not just charts
– let’s say title, let’s talk about headers and footers, let’s talk
about logo placement – each and every object in your
dashboard should have at least a .5 centimetre to 1
centimetre gap between each object. I think Tableau 10.4,
the latest version of Tableau, is quite concentrating on this
stuff which is quite exciting. I’m very eager to try that as
well. So they’re going to come up with a concept called
‘spacing between objects.’ If you set 0.75 centimetres as to
spacing, all the objects will be evenly distributed and will
have a white space around that.
Things like that would let the user clearly differentiate
various objects and concentrate. And your logo placement
and colours you choose, that’s very important. When it
comes to visualization, colours are very important. Also, I
know that you have a separate course in Udemy for colours
alone.
Kirill: Yeah, yeah. I personally learned so much from creating that
course. I didn’t realize how colours are important. I mean, I
realized that they’re important, but I didn’t realize how
exciting this theory of colours is. Yeah, you’re right, it’s
really cool.
Deepak: Yeah. And when it comes to colours, you have to make sure
colour blind safe. And when you are very sure that you have
a brand and you wanted to generate a colour palette from
the brand logo, you have certain tools in website like Coolors
– coolors.co – that’s a very good website for you. If you just
upload your picture, it will give you a colour palette to follow
so that your colour palette, whatever you are following in the
dashboard, would almost go with your company standards.
Kirill: Yeah, and there’s a couple of other tools. There’s a Coolor at
Adobe – color.adobe.com – the same thing where you can
upload your image. Yeah, Adobe Coolor, maybe that’s the
same one we’re talking about. But you can find a couple of
those tools online, they’re very useful, I agree. Sorry, you
were going to say the next thing?
Deepak: Yeah, things like this would be — you have to leverage many
things when you come up with a dashboard. Not only data
tells the story. Things like colours, how you maintain the
real estate of the dashboard, and titles, how good is your
title. I can tell you that just from your title, people get 30 to
40% of what is there in the dashboard. You have to very
carefully choose your title. If you are talking about sales and
if you return something like ‘Profit Margin’ as your title,
people are going to look for something called ‘Profityou’re
your dashboard and there will not be any value for profit
and they’re definitely not going to like it.
And the final one would be the data quality. The data quality
has to be checked not only once or twice, at least thrice,
because when you put in the dashboard, it is like you’re
certifying that this is the dashboard and this is where you
have to derive insights from. And just think about the
scenario where you have sourced the data from the wrong
place and people are making decisions out of that
dashboard. So you have to be very sure of what data you’re
using. That’s very important as well.
Kirill: That’s really cool. Let’s sum it up, these insights. We have
the data-ink ratio, which is actually a cool term. Data-ink
ratio should be increased, should be quite high, meaning
that it shouldn’t have less stuff that doesn’t convey insights
that are auxiliary or they’re just there for visual help. You
should make sure the space between objects is sufficient so
that it all looks good, it doesn’t look like it’s cluttering each
other. Colours are very important, and we have some tools
such as — you just sent me Coolors, so coolors.co or you’ve
got colors.adobe.com. It can help you derive the colours and
make sure it’s colour blind safe. Then we’ve got the title, so
we have to have a good title and data quality.
I like how you put it, that data quality has to be checked at
least three times. You are basically certifying that it’s all
correct and this is where you derive data. I have an
interesting question for you. A lot of times I hear from data
scientists about pie charts, that pie charts are evil and they
should not be used. What’s your opinion of pie charts
because I’ve seen them used quite a lot in BI?
Deepak: There’s also a very good blogger, I’m sure you know him, his
name is Andy Kriebel.
Kirill: Yes, yes. It’s so funny, he’s coming on the podcast very soon.
Deepak: Yeah. So, his tagline is “Friends don’t let friends use pie
charts.” (Laughs) That’s important. Pie chart has its own
way of expressing the information. In my experience, I would
use pie charts in only one instance, which is individual to
whole comparison. Let’s say you wanted to compare what
are all the mobile providers. For example, let’s take Apple,
Samsung, LG, Nokia and so on, and let’s say you want to see
what is Apple’s share and others. Let’s say that Apple will be
35% and rest of all will fall into 65%. That is called
individual to whole comparison. For those alone, I would use
a pie chart. Other than that, if you have more than three
dimensions, I would not suggest a pie chart. Other than
individual to whole, I would not suggest a pie chart
anywhere else.
Kirill: Okay. That’s really cool. I will bring this up with Andy
Kriebel, talk about it more with him. This is an interesting
point. Thanks for sharing that. And what’s your favourite
type of chart?
Deepak: When it comes to charts that are simple, I like the bar chart.
But when it comes to complicated stuff, when you want to
see the distribution, I like box plot more.
Kirill: Okay. Why do you like box plot so much?
Deepak: Box plot tells you a different story. Every time you click on a
box plot chart, it tells you a different story.
Kirill: Okay. By the way, these visualizations, this is the next thing
I want to talk about. These visualizations aren’t just
visualizations. They’re called dashboards for a reason,
because they’re interactive. You create them in a way that
people can click, so when they click on something, the rest
of the dashboard adjusts and talks about that specific part.
For instance, if you have a bar chart for different age
categories, or different age groups of your users or age
groups of your customers, and then the user of the
dashboard click on everybody on the bar that’s 25-35, then
the rest of the dashboard will focus in on everybody who is
25-35.
Tell us a bit more about interactivity. How important is it for
these dashboards to be interactive, and have you noticed
any changes in the behaviour of users of these dashboards
over the recent years? Because it’s becoming more and more
easier to create dashboards which are interactive, I think
people are getting more used to them now.
Deepak: When we talk about interactivity, if you ask me, interactivity
is nothing but asking questions. When we talk about asking
questions, we wanted a technology or a product to remember
what we asked before the second question. For example, I
asked three questions and we wanted the third question to
be answered after considering the first two questions I
asked. This is the power, right? So, when you ask one
question—for example, if I ask the Siri app — I’m talking
about the Siri voice command platform in Apple — if I ask
Siri what is the weather in Sydney, it tells me what is the
weather. And the second question is, should I bring an
umbrella today. So it should consider the weather, which is
my first question, and it should answer the second one also.
So, the correlation is what I’m talking about. This is what I
consider is very important when it comes to Qlik and
Tableau. Also, a very favourite question – I don’t know
whether you would ask it – but I would put it in the podcast:
Everyone always asks me which is better, Qlik or Tableau.
Kirill: Oh, yeah, that’s a good one. So what’s better?
Deepak: That is one question that everyone asks me whether it is in
the interview or whoever I meet and talk about the
visualization tools. According to me, Qlik has its own power
and Tableau has its own power. For Tableau, its power
comes from the drag and drop interface, very intuitive,
people can come up with a dashboard in just two to three
minutes and things like that. When it comes to Qlik—
Kirill: When you’re comparing Tableau versus Qlik, is it QlikView
or Qlik Sense?
Deepak: I would compare — if it is going to be a head-to-head
comparison, I would take Qlik Sense because it follows the
same methodology as well: drag and drop, easy, intuitive.
When it comes to Qlik, the power lies behind what I call a
‘subscripting layer’ where you can blend the data, where you
can play with the data, multiple layers of ETL.
For example, you have [indecipherable], you wanted to make
it as a mail, and you have a particular number you want to
add it up with another number and do some mathematical
expression layer by layer, Qlik is very good at doing it.
Tableau to an extent, yes, but not like Qlik. Qlik, when it
comes to dashboard building and publishing and getting it
to the user, it follows its own pathway, Tableau follows its
own pathway. But the power of Qlik lies in associative
experience.
The associative experience, if you ask me to talk, I will talk
for days and hours. It’s a very beautifully patented one that’s
patented by Qlik. That’s the one thing that will pull a big
crowd into Qlik. Tableau pulls the crowd because of its
intuitiveness and drag and drop interface.
Kirill: Which one would you say is easier to learn for somebody
listening to this podcast who wants to start?
Deepak: That depends. For someone who is listening to the podcast, I
would say—I have a hard and fast rule. If your data lake is
very good, if you feel that your data doesn’t need any
massaging at all, it is already done and if you feel your data
is very good, go for Tableau. If your data needs multiple
levels of cleaning, massaging, and you don’t have a data
warehouse at all, you don’t have a massaging layer in your
enterprise at all, Qlik comes to rescue you. So you directly
connect to an interface, you clean, massage and save
everything in QlikView data format and then you consume it
to the presentation layer. That’s the way.
Kirill: Cool, gotcha. And you mentioned as well, you work with
Power BI. What do you think of Power BI? I ask this question
because I’ve noticed them—it’s a Microsoft product, for those
who don’t know, and it’s free as opposed to Qlik and
Tableau, it has the same goals of what it can do—but what
I’ve noticed is that it has been developing and growing with
huge leaps every single month for the past year or so.
They’ve been releasing updates every month. And it stands
to show, a proof of their growth is how the Gartner’s Magic
Quadrant has positioned them a year ago.
Gartner’s Magic Quadrant is for all products and they have
one for BI products, it comes out in February, and last year
they were somewhere in the middle among everybody else,
but now they’re in the lead alongside Tableau in that Magic
Quadrant. So what are your thoughts and views of Power
BI?
Deepak: For me, Power BI stands great because every two weeks they
give proper updates. And the second thing is the integration
point of view. People want flexibility now. If you say, “This is
the way my product works,” and you don’t get more
flexibility, then people basically don’t like your product. Just
think about Power BI. It can integrate to any Microsoft
products seamlessly. And imagine embedding your
dashboard in your SharePoint. Take any organization. 60%-
70% of our organization is all Microsoft-based. I have
SharePoint in my organization; I have Excel, Microsoft Word,
PowerPoint laying everywhere.
So whenever you talk about BI tools to people, the first
question, the first good question they ask is, “Can I export it
to Excel? Can I export it to PPT?” because that’s what is
going to be shown to all the board members and CXO level
executives. So, when it comes to this, Power BI understand
these things and they export functionality and connecting to
all its database SQL, SSIS, SSRS layers, and DAX cubes. It
clearly understood what it wants. And also, for anyone who
knows Excel, catching up with Power BI is very easy. Just
imagine the number of people who use Excel in this world.
And if you say that will be a very less learning curve if you
know Excel, then who wouldn’t learn Power BI?
That’s the point. It’s about grabbing your share perfectly.
Microsoft did that perfectly, I would say. It clearly
understood who all can be its end user and what other
words can attract people like ‘Excel users’ and ‘Microsoft-
based organizations.’ They can use these words to easily pull
people inside. They’ve clearly concentrated in those areas
like exporting functionalities and connecting seamlessly into
Microsoft and being on cloud, which is a buzzword now
everywhere. If you say, “You can go on cloud in just five days
with our product,” then definitely you get one step ahead
than any other product.
Kirill: Yeah, totally. That’s a pretty solid overview. It will be
interesting to see how all these three compete and other
products compete over the next couple of years. It’s kind of
like a horse race. Every year you check who’s in the front,
who’s in the lead, who has got what new. Yeah, exciting
times.
Deepak: Yeah, did you see that? Yesterday, or I think day before
yesterday, Tableau acquired a company called ClearGraph.
Kirill: No, I didn’t see that. That’s interesting.
Deepak: Yeah. ClearGraph is a company—say you just type ‘Sales by
country’ and hit enter, it will give you a ‘Sales by country’
bar chart just by your search.
Kirill: Nice. That’s really cool.
Deepak: They’re doing a lot of things in natural language, NLP and
text processing. So, the ClearGraph acquisition definitely
shows that all the BI tools’ next mission would be
understanding whatever people type, like Google. So, the
Google-like experience is going to be the next big thing in the
data visualization world.
Kirill: That’s really cool. Then the only think you have to watch is
the correctness of it, right? It’s like is it actually giving you
what you want or something random.
Deepak: Yeah.
Kirill: And what I also like about Tableau is that they have a very
strong user community. There’s lots of bloggers who write
about Tableau, there’s lots of people on the forums, so if you
have any questions you get them answered very quickly. I
don’t think that’s the case for QlikView and Power BI, not yet
anyway. What do you think of that?
Deepak: True, very true. The community and the way they market
their product and the conferences that are being held every
year and the way they promote the conference, I think they
have very good team structure. One team promotes a
community, one team answers all the queries of people, and
the learning community is almost free. You can learn from
zero to hundred, everything for free when it comes to
Tableau.
I accept some mastery tricks that you provide in your Udemy
course, I would say. Other than that, if you want to just
understand the basics and just keep going until even
intermediate/advanced, everything is properly covered in
their course curriculum for the very basic users. But to go to
an advanced level, you need guidance from people who
achieved that already.
Kirill: Yeah, I totally agree with you. I’m not going to argue with
that. I think you can totally get everything for free in the
Tableau courses and online on YouTube. The only kind of
difference and what my courses bring is the element of a
journey and the elements of more applied case studies that
you go through and you practice and so on. It’s kind of a
different approach to education, but yeah, I totally
appreciate and it’s really cool that Tableau provides this
education and that people can pick it up pretty quickly.
Good thing.
I’m cautious of time, we’ve talked a lot about these different
tools and business intelligence. It’s so great to get some of
your insights. I’ve already learned a lot personally. And I
would also like to talk about your personal journey as well.
From your LinkedIn I can see that you only recently joined
ASG Group. How are enjoying it at the company? Do you
enjoy the company culture?
Deepak: Yeah, the company culture is really enthusiastic, energetic,
and more than that. I heard one important statement from
my boss which is listening to people, listening to our
customers. So, when you listen to people, instead of just
talking to them, listening is the one important capability
that every BI consultant should have, which was a good
learning for me when I started this job. When you put
yourself in that perspective, when you listen to people
instead of just pushing whatever you have in your mind, you
will just open up to give more suggestions.
Instead of giving an idea of what a business intelligence
layer does, you will give them an idea that “Your data
warehouse itself has a problem. Do you mind if I take a
look?” Things like that. So if you ask those questions, then
people open up. “Okay, do you want to do some consulting
on data warehouse and give us some insights to what’s
wrong with our data warehouse?” and then let’s go to the
next layer, which is business intelligence. That’s when it has
to start.
So I strongly feel that instead of just selling your single
product, it has to be an adaptation of their very own data
culture. Instead of you just selling, you have to adopt their
data culture and provide an end-to-end data solution. That’s
what ASG is all about and that’s all the services we provide
to our customers. And that’s why people keep returning to
ASG. If you just trace back any people who have worked
with ASG, they would come back again just because of this
quality.
Kirill: That’s fantastic. And I really like what you mentioned about
listening. I think that’s very important. You can save a lot of
time and efforts if you just listen carefully to personal
requirement, and not necessarily just in consulting. Even if
you’re within an organization trying to help somebody, if you
listen to them carefully and not make assumptions about
what they want and what their problems actually are, if you
hear them out you can really save a lot of time by focusing
on the right things.
I’m looking at your LinkedIn again. Throughout your career,
you’ve gone through a huge transformation, a huge growth.
You started as an Assistant System Engineer back in India
and then you had some big roles like Consultant, Associate
Manager, BI Specialist and Senior Technical Consultant for
several years at a time. You obviously have a vision for your
career. You’ve been, as I can see, deliberately selecting these
opportunities and growing and growing.
Now you’re in Sydney, Australia, working for ASG Group and
it sounds like you’re having lots of fun. What would you say
have been the most important qualities for you that helped
you build this career for yourself?
Deepak: What I feel is, from my academics, when I study a topic, I
will not turn my pages until I’m very clear about the topic.
That put me apart from the crowd. During my college or
school days, if you see my answer paper and other answer
papers in my class, people would have answered for 100
mark and score 98 or 99 or even 60 or 50, and if you take
my answer paper, I would have only answered for 60 mark
and I would get all 60. That is one characteristic which has
always put me outside, as an outlier from the total crowd.
So, the way I study is, until I understand a certain topic
completely, I will not move further. Let it be about BI or a
data warehouse or any other topic, if I’m reading about the
topic and if I’m stuck in the second line and if I’m not
understanding the word from the second line, I’ll just Google
about it, read about it, get myself clear, and then I will start
reading the third line. I will not cross the second line until
I’m very clear. That’s one good quality I have, which is
attention to detail.
The way I achieve mastery in any skills I take—for example, I
started my career, as I told you, with Unix, then Oracle, then
Informatica, then Cognos, then MicroStrategy, then
QlikView, then Tableau. So what I feel is an important
quality is your basics, no matter what, always have to be
very strong. Your advanced quality in certain tools or
products can be wavering, but your basics of any product
should be very strong. If your basics are very strong, for you
to reach mastery level is always a cakewalk. But if your
basic itself is not strong, although by mistake or by luck you
reach mastery level, it’s very highly likely that you might fall
down on any day. So I strongly feel if one’s basics are figured
out clearly, then you are very good to go.
Kirill: Fantastic. That’s really cool. On that first one, just to recap,
first one was a very unique quality. I don’t think I’ve
encountered this in person like this, where you will just
focus on that one thing that you’re currently working on and
you will make sure you know it really well and then you only
move forward, and the second one was basics have to be
very strong.
On that first one, I wanted to ask you, have you ever
encountered a situation where you just weren’t able to move
forward, where you came into a roadblock and it was just
impossible to either learn that thing or get that thing done or
whatever you were focused on, there was something
preventing you from doing it? Has that ever happened and
what did you do in that case?
Deepak: Luckily, I’ve not come across this situation. The reason
being is we are wealthy enough as a world, we are wealthy
enough that we have excellent resources available and tools
like Google, where you just type and it gives you whatever
we have. We are not here to invent something new.
Everything is there available on the Internet. For me, at
least, I think if I would have been born before the 1940s, I
would have come across this.
Kirill: Yeah, gotcha. Good point. With Google, you can pretty much
find an answer to anything. Okay, that’s been really cool.
You know, we’re slowly coming to the end of the podcast and
I wanted to ask you, from what you’ve seen in the space of
business intelligence and from the work you’ve done, where
do you think this whole field of data science is going and
what should our listeners look into to prepare for the future?
Deepak: When it comes to the future, the future doesn’t want people
who just have a balance of technology. The future needs
people who understand the technology, who understand the
process, who understand people and deliver something
which balances everything together. We wanted someone
who can listen to the problems and deliver solutions,
custom-tailored solutions according to the need, not the
same one-size-fits-all. It’s old theory now. Everyone needs a
different stitch and everyone wants it in their own way and
you have to make sure you have got the right skills to
provide that.
Kirill: Gotcha. People, process, technology – those are three
important things. Can you elaborate on that a little bit
more?
Deepak: When you want to sell or project some idea about yourself or
any future-proof product you want to come up with, just
imagine your product with certain end users. Any product
that has been built or going to be built without keeping
some end users in mind, the product will definitely not
succeed. And same with the process.
For example, I’m going to come up with a product called a
newspaper. Let’s say that the whole world doesn’t have
something called newspaper for now. If I say that I’m going
to come up with a newspaper now, how many people would
really read a newspaper when all the information they want
is available across the Internet for free? Although you charge
1 cent, do you think people would buy that? No. So when
you come up with a product, or when you come up with a
solution, or any platform or a service or anything, you have
to be very sure of who are going to be your end users and
who are you going to serve and what is the current problem
they are coming across and how are you going to solve it? So
these three things are all I’m talking about: people, process
and technology. It has to be perfectly balanced.
Kirill: Fantastic. Thank you so much, Deepak. That’s really good
insight. We’re going to end on that note. Thank you so much
for coming on the show. A quick question: Where can our
listeners find you, connect with you or get in touch so they
can follow your career further?
Deepak: They can find me at LinkedIn. They can also reach me as a
consultant through ASG.
Kirill: Gotcha. ASG in Sydney and LinkedIn.
Deepak: Yeah, LinkedIn. Just search for Deepak Prasad.
Kirill: We’ll definitely share that. And I have one last question for
you. What is a book that you can recommend to our
listeners that can help them become better data scientists?
Deepak: If you want to become a data scientist, first you have to have
a really good hold of your data. When you don’t have a good
hold of your data, no matter what algorithms you know or
what skills you get for data science, it will go for a toss. So
for data analysis purposes, I strongly recommend a book
“Head First Data Analysis” by O’Reilly. And the second book
I suggest is a book called “Data Smart.” That is for data
analysis as well.
For visualization, I strongly feel there is a very good book
that I will read again and again and again that is called “The
Truthful Art” and “The Functional Art.” There can’t be any
better book than this according to me. And I bought a good
book recently which is called “The Big Book of Dashboards”
which is quite interesting as well.
And for data science—for me, when it comes to data science,
you have to have a good hold of many algorithms and real
use case studies. That will basically come from the “Data
Smart” book itself, but there are many other data science
books called “Doing Data Science” from O’Reilly and that’s a
good place to start. And then once you have decided you’re
going into data science, “Doing Data Science” is a better
start, and then once you’ve chosen which technology you’re
going to take, whether it’s going to be R or Python, according
to that, you have to choose your books.
Kirill: Wow. That’s a whole library there, guys. “Head First Data
Analysis,” “Doing Data Science,” “Data Smart,” “The Truthful
Art,” and “The Big Book of Dashboards.” I actually heard of
that last one, “The Big Book of Dashboards.” I haven’t read it
yet, but somebody on the podcast mentioned it so it must be
pretty good. I’m sure you’re going to enjoy that one. Once
again, thanks a lot, Deepak, for coming on the show. I really
appreciate you coming in and sharing all these insights. It
was very interesting to talk about business intelligence and
different types of tools and approaches, so thanks so much.
Deepak: Thank you, Kirill. Thanks a lot for this opportunity. And
more than me talking, I learned a lot from you as well.
Kirill: Thanks, man. All right, take care.
Deepak: Thanks. Take care. Bye.
Kirill: So there you have it. That was Deepak Prasad from ASG
Group. I really hope you enjoyed today’s episode and that
you picked up some interesting ideas from here, whether it
be about business intelligence or about your career. And
personally, my favourite part was when we discussed the
three pillars that you need to take into account when
deploying business intelligence tools. That is people, process
and technology. And personally, I actually think that it’s not
just relevant to business intelligence, but in many areas of
business, you need to think about these three things, about
people and process and technology and how they all work
together, how people use the technology to execute the
processes that they should be executing, how the processes
are tailored to the people to maximize the power of
technology, or how the technology is selected to help the
people execute the processes. So there’s lots of ways that
these three interact and depending on the current business
situation, it’s important to look at them from different
perspectives through different lenses.
So there you go. You can get the show notes for this episode
at www.superdatascience.com/87, and of course, there you
will find all of the resources mentioned in today’s show,
including a URL to Deepak’s LinkedIn where you can hit him
up and connect with him and see how his career progresses
further. And I look forward to seeing you here next time.
Until then, happy analyzing.