the

Data Science
CAREER GUIDE
The Data Scientist
Some years ago, a completely new position appeared on the horizon –

the data scientist. It was little-known and extremely mysterious as there

was no ‘data’ on it. Quite quickly the laws of economics started governing

the recruitment and the scarcity of supply drove up the price of data

scientists. According to Glassdoor, since 2016, data science has been the

best career to pursue.

The Data Scientist position is absolutely fascinating due to the variety of

activities undertaken and the expertise needed to perform the job.

In this guide, we show you the different career paths you can take if you

want to end up in the Data Scientist position.

The Rise of Data Science
Data science truly differentiated itself as a unique field with the
emergence of the first Data Scientists. The pioneers in this field were
people ahead of their time with knowledge in multiple disciplines and
incredible understanding of actual business processes.

Nowadays, we have sophisticated software at our disposal, such as

Google Analytics, Tableau, Power BI, even the performance revamped
Microsoft Office. There has also been significant development in
programming languages like Python and R, which are easily customized
for specific activities. Furthermore, statistics has thrived due to the
increased computational power. Lastly, as machine learning went out of
academia and into the real-world, business changed forever.
Which industries use data?
Hint: which industry doesn’t?
Aerospace
Industry uses
lots of data

Logistics? Data, data, data

Automobile Industry uses data

Every digital
device from
your car to
your watch..

Sometimes bad data is the

most valuable data…

NYSE thrives on data

 Nutrition

Medicine

Even baseball and farming!

What does it mean to be in
the business of data?
The 365 Data Science BI Analyst Data Analyst Data Scientist

Career Tracks Mathematics

Probability

Microsoft Excel
This table summarizes the career Statistics

tracks and their required Intro to Data and Data Science

competencies. In this guide, you R

Python
will find detailed information about
SQL
the positions and the courses that
Tableau
you may pursue in order to land a
Advanced Statistics
job in the industry.
SQL + Tableau

SQL + Tableau + Python

Machine Learning
Opportunities
Data science presents many opportunities for people who are We at 365 DataScience have prepared this guide for

quantitively inclined, some more than others. Either way, one can easily three different data science careers that you may want to
pursue.
imagine how some data science is, or will be, required for positions like:

✓ Marketing Analyst;

✓ Business Analyst;

✓ Data Analyst;

✓ BI Analyst;

✓ Data Scientist

Of course, each of those requires relevant background – either

appropriate education, or work experience. Unfortunately, few

institutions manage to provide sufficient and relevant preparation.

That’s where online courses step in to fill the gap.

The Business Intelligence Analyst
The Business Intelligence Analyst Track
The business intelligence branch of companies is one of the hottest topics in
recent years. It is one of the jobs that job seekers understand but sometimes
deem out of reach because of the specialized skills required.

Let’s first differentiate between a business analyst and a business intelligence

analyst. While the business analyst works with data and makes data-driven
decisions, a BI analyst is able to perform much more technical analyses.
Executed based on larger datasets where competence with specialized software
(such as Tableau, Power BI, etc.) is a must. Additionally, SQL programming is a
great advantage.

Some of the main duties of a BI analyst are gathering data, preprocessing

databases, market research, trend analysis, data visualization, reporting,
dashboard creation, and making recommendations.
The Day of the BI Analyst

The BI analyst has two defining traits: they work with inhouse data and have a
business orientation. These also define the two main skillsets needed - data and
business.

To be more specific, let’s say you have been tasked with preparing a report about
how long computers have been on in the office (uptime). If you are the first to
ever do this, you will need to plan your data journey, design your metrics, gather
the data, and eventually analyze it. You will be expected to visualize it in a
manager-friendly way and tell the story of office computer uptime.

Becoming a BI analyst combines the worlds of business and data. All the skills
involved are easily transferable into other business or data science positions.
What is the required expertise for a
Business Intelligence Analyst?
The Business Intelligence Analyst

We have prepared a summary of the required skills for a BI Analyst, based on

the responsibilities that employers assign to the position.

The following list encompasses the main competencies that you may be
expected to have when entering a company. While it is highly recommended
that you are proficient in all of them, responsibilities vary from employer to
employer. Two BI Analysts could be asked to perform completely different
activities, even in the same department. This is a product of the specialization
of labor that is observed in the current economy.

No matter the particular job, you will be required to have at least conceptual
knowledge of these activities.
Expertise of a BI Analyst

1. Data
Data related responsibilities could be
arranged in four main categories:
Gather, Maintain, Design, and Analyze.
04 Processing and Analysis

03 Model (design) data and metadata and

manipulate variables

02 Primary support, maintenance,

and quality assurance

01 ETL, data mining, mapping

and retrieval
Expertise of a BI Analyst

2. Business
Business related responsibilities could also
be arranged in four categories: Identify,
Evaluate, Report & Recommend, and
04 Optimize staffing levels, personnel responsibilities;
automate processes
Optimize
03
Prepare financial and operational reports.
Visualize using rich and dynamic
dashboards. Tell a story and recommend

02 Performance, KPIs, user behavior

01 Problems, root causes,

risks, primary insights
So, how should you approach a
Business Intelligence Analyst career?
Landing a BI Analyst job depends on these skills

The responsibilities of a BI Analyst may vary, but

95% of the time, you will be using one of these 6
skills. You should be extremely familiar, if not
Intro to Data and DS proficient, with all data science related terms and
concepts, Microsoft Excel, Statistics fundamentals,
Microsoft Excel
and a business intelligence tool, like Tableau.
Statistics Knowledge of SQL is not required for all BI
positions, but is a great plus. Furthermore, being
Tableau
able to integrate SQL with Tableau is a skill that will
SQL truly differentiate you from the other applicants.

SQL + Tableau
1. Intro to Data and
Data Science
The best way to start exploring a
position in data science is with a
comprehensive introductory guide
such as this one. However, a better
alternative is to take a course which
aims to summarize, organize, and
explain all data science buzzwords,
terms, tools, approaches, and
techniques. Only after you have seen
the bigger picture can you put the
pieces of the puzzle together and dive
into studying data science.
1. Intro to Data and Data Science
The image on the right can help us gain an idea
of the relationship between different fields in
data science. Moreover, it provides examples of
real-world activities related to each data science
field.

Business analytics, data analytics, business

intelligence, machine learning – while similar,
they are far from the same. Their intersection
is data science.

Logically, becoming a good BI analyst, data

analyst or data scientist requires you to be able
to classify each problem to its related field in
order to apply the appropriate techniques.
2. Microsoft Excel
Microsoft Excel is a powerful software
and the most widely used
spreadsheet ever. Almost any job
nowadays features Excel and being
truly proficient at it has become a
must. Combined with the power of
different plug-ins, you can customize
this software to become more useful
for just about anything – from
statistics to word processing. While
little known, a lot of number
crunching (especially for smaller data
science projects) is done in Excel.
 2. Microsoft Excel
• Sales data
• Markets
• Experiments • Consumer behavior
Design Analyze
• Your data
• Custom metrics

Identify
Perform
and report
• Customize statistical tests • Visualize your data

• Develop tools and data models • Trends, patterns, anomalies

• Create reports
3. Statistics
Statistics is the basis of all data
analytics. It is paramount that an
analyst understands the roots of the
tests performed in order to interpret
them. You should be comfortable with
the concepts and how to implement
them into tests and experiments.
Sometimes, analysts are expected to
suggest metrics to be measured and
experience with statistics is the right
way to approach such problems.
 Hypothesis A/B testing

3. Statistics
ANOVA testing

Margin of
error Central limit
At the workplace, a BI Analyst is expected theorem
to understand the root of various
problems. She should be able to rapidly
Basic
identify possible reasons for both under-
and overperformance of certain metrics.
Statistical
significance
concepts
Correlation
While business judgement is needed,
data-driven decisions are formed
through statistical tests.
p-value

Distributions

Regression
Confidence
analysis
intervals
4. SQL
SQL is a domain-specific language that
serves the niche of relational database
management. It is mandatory for
anyone employed in data science to be
able to work with databases and SQL is
the way to go. There are different
platforms for SQL, such as Oracle,
MySQL, and Microsoft SQL Server.
While they have their own peculiarities,
the underlying language is virtually the
same.
 4. SQL
At the workplace, one often needs information from the
database. There are two options: extract it on your own,
or contact the IT team. When you are the BI Analyst, you
usually need all data at all times and don’t want to
depend on another person. Apart from utility, it is also
the responsibility of a BI
to interact with a database
and pull whatever is needed Export
for her data-driven decision. Transform
Load
Design
5. Tableau
The best description of Tableau comes
from its creators: ‘Tableau can help
anyone see and understand their data’. It
is the leading visualization software in
the business intelligence and data
analytics field in the recent years.
Whenever you see beautifully visualized
data, chances are that Tableau has
something to do with it.

Certainly, other BI tools do exist, such as

Power BI and IBM Cognos, however,
Tableau is the most popular one.
5. Tableau
Working with Tableau automatically gives you a competitive advantage as it helps
you navigate through and understand massive amounts of data in seconds

Visualize data with customized tools Analyze KPIs with fresh eyes after
for just about any purpose. Report seeing what your data actually means
by sales, location, focus group, and and present it in the most engaging
much more. way.

Perform meaningful breakdowns on Increase client engagement and

any dimension of your data and conversion rates through insights
easily uncover hidden gems. about brand awareness, trends,
patterns, and anomalies.
6. SQL + Tableau
Knowledge of SQL and Tableau are two
indispensable skills for a BI analyst. What
truly distinguishes an analyst from his/her
peers is interdisciplinary knowledge,
breadth, and ability to combine expertise
from different domains. One of the most
impressive ways you can differentiate
yourself from other analysts is the ability
to work with data from the very source
and then present it through beautiful,
meaningful and professional
visualizations.
6. SQL + Tableau

Plan Connect with SQL Preprocess Visualize

• Define the problem • Access the database in • Design KPIs • Plot the data
• Figure out how to Tableau • Preprocess the data in • Visualize professionally
acquire the data • Find all relevant calculated fields • Create dashboards
• Think about information
visualization
SQL

Plan your data journey. Once you connect your database Raw data is rarely suitable for It is up to you to create the
Professional visualization implies with Tableau, you can write queries visualization right away. More often visualizations intended in the beginning.
thinking your problem through to extract information directly in the you must use some degree of Tableau provides a seamless experience
from data collection to the axes interface. Additionally, you can preprocessing to design the metrics both for single plots and professional
of your final plots. combine several databases. you’ll later visualize. dashboards.
FAQ at interviews
1. Describe the different parts of an SQL query.
2. What is the difference between INNER JOIN and OUTER JOIN?
3. You have a table called with Cust_ID, Order_Date, Order_ID, Tran_Amt. How would
you select the top 100 customers with the highest spend over a year long period?
4. If you were stuck on a desert island with a database that contained all the
knowledge ever created, but you only had 10 SQL statements that you could ever
use, what would they be?
5. What is the difference between DELETE and TRUNCATE? What is the difference
between UNION and UNION ALL? What is the difference between a WHERE
statement and a HAVING clause?
6. The conversion rate for a specific chair is 0.5% for the first 50,000 shoppers that look
at it. The price of the chair is $250. Our company makes 27% profit on the sale. The
next 50,000 shoppers will get a 10% discount. What is the conversion rate we must
achieve to receive the same profits as before?
FAQ at interviews

7. What experience do you have with Tableau? Our BI team is brand new and is
under-financed. We have no standard procedures or training and everything
is ad-hoc. How would you go about this situation?
8. You get X amount of views on a website, Y amount of people click on the ad,
then Z amount of people enter their names after, where X, Y and Z are given.
How much does it cost to acquire a customer? What’s the conversion rate?
Would it make sense to run the campaign comparing the value of customer
acquisition to the revenue gained from conversion rate.
9. You have been asked to send an e-mail campaign to customers that have
made a purchase on Amazon.com in the past but not recently. How you
would go about the process. What query would you use?
FAQ at interviews
10. Our firm is going to send 2 different catalogs to their customers. One of the
catalogs costs 50 cents to make and is 50 pages long. The conversion rate for
the catalog is 5% and each customer brings in 315 dollars. The second catalog
costs 95 cents to make, is 100 pages long and each customer brings in 300
dollars from it. The profit margin is 30%. What should the conversion rate for
the second catalog be to make at least the same amount of profit as the first
one. After you find the conversion rate for the second one, there is a second
part of the problem. Wayfair is planning to make a new catalog which is going
to cost 10 cents more than the 100 page one. The more expensive catalog is
going to be sent out to 20% of the customers while the remaining 80% are
going to get the 100 page one. Assume the same 30% profit margin and $300
profit from each customer. What should the conversion rate for the new
catalog be in order to receive the same profit at the end?
The Data Analyst
The Data Analyst Track
The data science department of companies is the most rapidly growing one in recent
years. The data analyst is the building block of a data science team. While more and
more individuals learn about the position, many still do not understand the nature of
the work or simply don’t have the skills to perform the job of the Data Analyst.

The Data Analyst is similar to the BI Analyst. While a BI analyst performs technical
analyses based on large datasets, the data analyst creates and runs complicated
statistical models to not only extract insights but also predict outcomes. Ideally, the
Data Analyst has deep statistical knowledge and superior programming skills; this
makes her much more capable than the BI Analyst to work with big data. However,
less business knowledge is needed to be a Data Analyst – it’s actually all about the
data.

Main functions of the Data Analyst are gathering data, creating and running models,
trend analysis, testing, visualizing, making recommendations, and storytelling.
The Day of the Data Analyst

There are three major activities for a Data Analyst: data cleansing & management, programming &

analysis, and presentation of findings.

Let’s say that you are a Data Analyst and you are asked to create a model which identifies units that

are likely to become faulty. How would you typically approach the task? First, you would get your

data, or design a way in which you can gather data in a given timeframe. Then you would create a

model that fits the observed dependencies (e.g. the more an item is used, the more likely it is to

break down). Once there, you’ll test your models and achieve a certain level of accuracy. Finally, you

would gather up the findings and create a presentation tailored to your audience. This usually

means manager-friendly and light on data and methodology, and you will explain what you found

to be drivers of the phenomenon.

Being a Data Analyst equals swimming in data. The more projects a Data Analyst has been through,

the deeper understanding of data analytics and predictive modelling in specific she’ll have and the

more valuable she will become to the employer.

What is the required expertise for a
Data Analyst?
The Data Analyst

We have prepared a summary of the required skills for a Data Analyst, based
on the responsibilities that employers assign to the position.

The following list comprises of the main competencies that you may be
expected to have when entering a company. While it is highly recommended
that you are proficient in all of them, responsibilities vary from employer to
employer. Two Data Analysts, even if they are sitting side by side, may be
asked to perform completely different tasks. This is a product of the
specialization of labor that is observed in the current economy.

No matter the particular job, you will be required to have at least conceptual
knowledge of these activities.
 Expertise of a Data Analyst
Filter, profile, assess third
party, acquire new,
control the quality of data
Design, implement, map, ETL,
Continually process raw data,
identify weak spots
test, optimize, move across
platforms, extract in different
formats

Data cleansing

& database management

Programming
Visualize and
Presentation of findings
tell the story
Analysis of the data of the data
Sophisticated statistical models

Analyze, find trends and patterns,

cluster, predict

Manipulate your variables, structure

tests, assess quality, find weak spots,
optimize models
So, how should you approach a
Data Analyst career?
Landing a Data Analyst job, depends on these skills

The responsibilities of a Data Analyst may vary, but

95% of the time, you will be using at least one of
these 7 skills. You should be extremely familiar, if not
Intro to Data and DS proficient, with all data science related terms and

Microsoft Excel concepts, Microsoft Excel, Statistics fundamentals,

and certainly a programming language such as R or
Statistics
Python. Most data analysts have expertise in both.
SQL Knowledge of SQL isn’t required but is always a plus.

R A Data Analyst truly shines with her knowledge of

advanced statistical methods, especially applied with
Python
R or Python.
Advanced Statistics
1. Intro to Data and
Data Science
The best way to start exploring a
position in data science is with a
comprehensive introductory guide
such as this one. However, a better
alternative is to take a course which
aims to summarize, organize, and
explain all data science buzzwords,
terms, tools, approaches, and
techniques. Only after you have seen
the bigger picture can you put the
pieces of the puzzle together and dive
into studying data science.
1. Intro to Data and Data Science
The image on the right can help us gain an idea
of the relationship between different fields in
data science. Moreover, it provides examples of
real-world activities related to each data science
field.

Business analytics, data analytics, business

intelligence, machine learning – while similar,
they are far from the same. Their intersection
is data science.

Logically, becoming a good BI analyst, data

analyst or data scientist requires you to be able
to classify each problem to its related field in
order to apply the appropriate techniques.
2. Microsoft Excel
Microsoft Excel is a powerful software
and the most widely used
spreadsheet ever. Almost any job
nowadays features Excel and being
truly proficient at it has become a
must. Combined with the power of
different plug-ins, you can customize
this software to become more useful
for just about anything – from
statistics to word processing. While
little known, a lot of number
crunching (especially for smaller data
science projects) is done in Excel.
 2. Microsoft Excel
• Sales data
• Markets
• Experiments • Consumer behavior
Design Analyze
• Your data
• Custom metrics

Identify
Perform
and report
• Customized statistical tests • Visualize your data

• Develop tools and data models • Trends, patterns, anomalies

• Create reports
3. Statistics
Statistics is the basis of all data
analytics. It is paramount that an
analyst understands the roots of the
tests performed in order to interpret
them. You should be comfortable with
the concepts and how to implement
them into tests and experiments.
Sometimes, analysts are expected to
suggest metrics to be measured and
experience with statistics is the right
way to approach such problems.
 Hypothesis A/B testing
ANOVA

3. Statistics
testing

Margin of
error Central limit
At the workplace, a Data Analyst is theorem

expected to understand the root causes

of various problems. She should be able
Basic
to rapidly identify possible reasons for
Statistical concepts
both under- and overperformance of significance
Correlation
certain metrics. While business
judgement is needed, data-driven
decisions are formed through statistical
p-value
tests.
Distributions

Regression
Confidence
analysis
intervals
4. SQL
SQL is a domain-specific language that
serves the niche of relational database
management. It is mandatory for
anyone employed in data science to be
able to work with databases and SQL is
the way to go. There are different
platforms for SQL, such as Oracle,
MySQL, and Microsoft SQL Server.
While they have their own peculiarities,
the underlying language is virtually the
same.
4. SQL
At the workplace, one often needs information from the
database. There are two options: extract it on your own,
or contact the IT team. When you are the Data Analyst
you usually need all data at all times and don’t want to
depend on another person. Apart from utility, it is also
the responsibility of a Data
Analyst to interact with a
database and pull whatever Export
is needed for her data- Transform
driven decision. Load
Design
5. R
R is a programming language specifically
designed for statistical analysis, data
manipulation, and graphics.

R’s readability, and vast library network are

key for R’s ability to rapidly and effectively
perform advanced analyses, including
creating machine learning algorithms.

Even if we step away from R’s speed, it’s

increasing popularity is by no means a fluke.
R is a programming language deeply rooted
in its user community: libraries, and task-
specific packages are constantly created and
improved by the users, with the size and
scope of R’s resources being second to none.
5. R VISUALIZATION
Immediately visualize your data with functions
R was designed specifically for data
tailored for any graphic you will need. Use
manipulation. That said, you can grammar of graphics alongside the ggplot2
perform the majority of operations you package to create meaningful, highly-
customizable visualizations, and plots.
need in data science – like filtering,
Think outside the common static canvas, and
arranging, factoring, gathering,
design interactive graphs for the web – R does
mutating, spreading your data, and this fast and effectively, too.
much more – in a handful of lines of
code.

BIG DATA
R is designed to handle extremely
big data sets, usually gathered by

COMPUTATIONAL medium to large companies, or for

ANALYSIS academic research.

Enjoy the full computational power of your

computer, exponentially increasing the
speed of the analysis through specialized
libraries and vectorized code.
6. Python

Python is an open-source, general-

purpose high-level programming
language. It is one of the most widely
used programming languages in the
past few years. The technical
advantages it has over other
programming languages and its
modules for scientific computing make
it a preferred choice while working in
the fields of in finance, econometrics,
economics, data science and machine
learning.
6. Python Leading BI software Tableau launched TabPy which is an integration of
BI Python within Tableau. In this way, you can create reports and
dashboards that leverage the real-time computational power of Python.

Python is compatible with MySQL and is expected

to be integrated in the new version of the Microsoft
DBMS SQL Server as well. Thus, giving you the capability of
working with relational databases. Furthermore,

Python Libraries Python can be used to produce non-relational

Easy syntax, open-source,

databases, such as NoSQL.
integration, scope
NumPy, pandas, matplotlib, Seaborn,
scikit-learn, and TensorFlow are some Python is the leading language used for
of the most widely used libraries in machine learning in the field of data science.
data science. They combine the Machine With the amazing scikit-learn and powerful
capabilities of many other
learning frameworks like TensorFlow, Keras, and
programming languages, but let you PyTorch, Python is easily the all-in one
use them all in one place in an programming language for ML.
environment that just needs to
support Python.
7. Advanced
Statistics
Advanced statistics in this framework
refers to the symbiosis between linear
algebra, computational power, and
predictive modeling. Examples are
linear regression, logistic regression,
and clustering. Depending on the
source, this part of data science can be
called anything from statistical methods
to machine learning. Either way, the
techniques involved are the same and
are pretty much what professionals
understand by the term ‘data science’.
 Math
7. Advanced Statistics methods Distributions

Bayesian
Transform inference
Basic statistics lays the foundation of the field and
focuses on frequentist inference. Advanced
statistics builds upon it, entering multi-
dimensional spaces, through knowledge of Advanced
mathematical methods, transformations and
Regression
topics
distributions. Moreover, more complex means of Decision
theory
analysis are introduced, such as regression,
classification, clustering and factoring. Finally,
Bayesian inference and decision theory allow the
Data Analyst to solve problems of dynamic and/or
behavioral nature. Factor Classification
Cluster analysis
analysis
FAQ at interviews
1. If you have a 10x10x10 cube, what is the outside surface area?
2. You have 10 bags with 10 stones each. One of the bags is lighter than the
others. Using a digital scale, how would you figure out which one is it with
just one weighting.
3. What is the sum of numbers from 1 to 100?
4. A snail falls down a well 50ft deep. Each day it climbs up 3ft and each night
falls down 1ft. How many days does it take him to get out?
5. How many SUV’s in the parking lot downstairs?
6. What is the difference between UNION and UNION ALL? What is the
difference between DELETE and TRUNCATE? How would you find median
value for a given columns?
7. Identify the issues in this excel spreadsheet.
FAQ at interviews

8. What kind of RDBMS software do you have experience with?

9. Draw a line that would give our company the same revenue $9 per sale, with
the y-axis (% On Time), and the x-axis ($ off for being late).
10. If a product costs $4.00, with a $800 sunk cost, and we charge X amount of
dollars along with a $10 annual fee, how many do we need to sell to break
even, etc).
11. Sales department increased the selling price of all items by 5%. There are 10
items, all with different price tags. Before the price increase, gross revenue
was $500,000 with an average selling price of $1. After the price increase
Gross revenue was $505,000, with an average selling price of $0.95. Why
hasn't the price increase had the desired impact of increasing revenue and
average selling price?
The Data Scientist
The Data Scientist Track
The data science department has been the most rapidly growing in recent years. Many
individuals do not know about the position, do not understand the nature of the work, or
simply don’t have the skills to perform the job of the Data Scientist.

The Data Scientist is on top of the data science ladder. However, describing her job gives
everyone a headache. In fact, the Data Scientist has such a slippery definition, that if you
look in five places, you will find five different definitions of what a Data Scientist is. For us at
365 Data Science, a Data Scientist is a person who has a broad range of knowledge in
multiple disciplines, while specialized in one or two. She understands the business processes
of a company, including marketing, strategy and sales, but also engineering and product
development. Nonetheless, where she truly shines is machine learning and statistics.

Main responsibilities of the Data Scientist are gathering data, structuring databases, creating
and running models & analyses; strategy, marketing, product placement, pricing, making
recommendations, and telling the story of the data.
The Day of the Data Scientist

One of the best definitions of a data scientist is:

‘A data scientist is a better statistician and economist than most programmers, a better programmer and
economist than most statisticians, and a better statistician and programmer than most economists’

An example: Think about a hotel chain. One of the most vital activities for them is revenue management.

There are two important considerations. First, some days are much more important for a hotel than

others. Second, customers are willing to pay several times higher prices depending on time of the year

and location. A data scientist can apply different statistical methods together with domain expertise to

identify the most important days. For the pricing part things are different. With machine learning in the

data scientist’ toolbox, she can predict with extreme accuracy the highest price customers are willing to

pay for a particular date and hotel. Best part? The whole process can be performed in real-time and

completely automated.

Given that the Data Scientist swims in data, and data is rarely super nice; she faces challenges at every

turn. But the more projects she goes through, the deeper her understanding of the business and machine

learning becomes, and the more valuable she is to any employer (or client).
What are the required expertise for a
Data Scientist?
The Data Scientist

We have prepared a summary of the required skills for a Data Scientist, based

on the responsibilities that employers assign to the position.

The following list comprises of the main competencies that you may be asked

to possess when joining a company. While it is highly recommended that you

are proficient in all of them, responsibilities vary from employer to employer.

Any two Data Scientists are different. This is because each one of them is

formed by her own experience, knowledge, and talent.

No matter the particular job, you will be required to have at least conceptual

knowledge of these activities.

Expertise of a Data Scientist

The Data Scientist is a complex professional that has broad expertise in

different topics, while specializing in statistics and machine learning.

She can do everything a Data Analyst can do

She can do everything a BI Analyst can do

Expertise of a Data Scientist

As the Data Scientist position includes everything

we’ve talked about so far, we will focus on its
specializations.

Advanced Statistics
01 Develop statistical models based on internal
and external variables, analyze using predictive
multidimensional analysis

Machine learning
02 Research, design, and execute algorithms for
numbers, text, emotions, images, decisions
and many more

Storytelling
03 Make a story out of the data & the machine
learning outcomes and tell it to people who
do not have technical knowledge
How should you approach a Data
Scientist career?
Landing a Data Scientist job, depends on these skills

The responsibilities of a Data Scientist may vary, but

you will surely be using one of these 11 skills. You
Intro to Data and DS
should be extremely familiar, if not proficient, with
Microsoft Excel
all data science related terms and concepts,
Statistics
Microsoft Excel, Statistics fundamentals, and a
Tableau
business intelligence tool, like Tableau. Knowledge
SQL of SQL preferable, but R and/or Python are a must.
R Furthermore, a data scientist should be able to
Python apply all these skills in advanced statistical methods
Advanced Statistics and machine learning. Integrations of SQL and

Machine learning Tableau, and SQL, Tableau, and Python will give a

SQL + Tableau data scientist an advantage in this competitive job

market.
SQL + Tableau + Python
1. Intro to Data and
Data Science
The best way to start exploring a
position in data science is with a
comprehensive introductory guide
such as this one. However, a better
alternative is to take a course which
aims to summarize, organize, and
explain all data science buzzwords,
terms, tools, approaches, and
techniques. Only after you have seen
the bigger picture can you put the
pieces of the puzzle together and dive
into studying data science.
1. Intro to Data and Data Science
The image on the right can help us gain an idea
of the relationship between different fields in
data science. Moreover, it provides examples of
real-world activities related to each data science
field.

Business analytics, data analytics, business

intelligence, machine learning – while similar,
they are far from the same. Their intersection
is data science.

Logically, becoming a good BI analyst, data

analyst or data scientist requires you to be able
to classify each problem to its related field in
order to apply the appropriate techniques.
2. Microsoft Excel
Microsoft Excel is a powerful software
and the most widely used
spreadsheet ever. Almost any job
nowadays features Excel and being
truly proficient at it has become a
must. Combined with the power of
different plug-ins, you can customize
this software to become more useful
for just about anything – from
statistics to word processing. While
little known, a lot of number
crunching (especially for smaller data
science projects) is done in Excel.
 2. Microsoft Excel
• Sales data
• Markets
• Experiments • Consumer behavior
Design Analyze
• Your data
• Custom metrics

Identify
Perform
and report
• Customized statistical tests • Visualize your data

• Develop tools and data models • Trends, patterns, anomalies

• Create reports
3. Statistics
Statistics is the basis of all data
analytics. It is paramount that an
analyst understands the roots of the
tests performed in order to interpret
them. You should be comfortable with
the concepts and how to implement
them into tests and experiments.
Sometimes, analysts are expected to
suggest metrics to be measured and
experience with statistics is the right
way to approach such problems.
 Hypothesis A/B testing
ANOVA

3. Statistics
testing

Margin of
error Central limit
At the workplace, a Data Scientist is theorem

expected to understand the root causes

of various problems. She should be able
Basic
to rapidly identify possible reasons for
Statistical concepts
both under- and overperformance of significance
Correlation
certain metrics. While business
judgement is needed, data-driven
decisions are formed through statistical
p-value
tests.
Distributions

Regression
Confidence
analysis
intervals
4. Tableau
The best description of Tableau comes
from its creators: ‘Tableau can help
anyone see and understand their data’. It
is the leading visualization software in
the business intelligence and data
analytics fields in the recent years.
Whenever you see beautifully visualized
data, chances are that Tableau has
something to do with it.

Certainly, other BI tools do exist, such as

Power BI and IBM Cognos, however,
Tableau is the most popular one.
4. Tableau
Working with Tableau automatically gives you a competitive advantage as it helps
you navigate through and understand massive amounts of data in seconds

Visualize data with customized tools Analyze KPIs with fresh eyes after
for just about any purpose. Report seeing what your data actually means
by sales, location, focus group, and and present it in the most engaging
much more. way.

Perform meaningful breakdowns on Increase client engagement and

any dimension of your data and conversion rates through insights
easily uncover hidden gems. about brand awareness, trends,
patterns, and anomalies.
5. SQL
SQL is a domain-specific language that
serves the niche of relational database
management. It is mandatory for
anyone employed in data science to be
able to work with databases and SQL is
the way to go. There are different
platforms for SQL, such as Oracle,
MySQL, and Microsoft SQL Server.
While they have their own peculiarities,
the underlying language is virtually the
same.
5. SQL
At the workplace, one often need information from the
database. There are two options: extract it on your own,
or contact the IT team. When you are the Data Scientist
you usually need all data at all times and don’t want to
depend on another person. Apart from utility, it is also
the responsibility of a Data
Scientist to interact with a
database and pull whatever Export
is needed for her data-driven Transform
decision. Load
Design
6. R
R is a programming language specifically
designed for statistical analysis, data
manipulation, and graphics.

R’s readability, and vast library network are

key for R’s ability to rapidly and effectively
perform advanced analyses, including
creating machine learning algorithms.

Even if we step away from R’s speed, it’s

increasing popularity is by no means a fluke.
R is a programming language deeply rooted
in its user community: libraries, and task-
specific packages are constantly created and
improved by the users, with the size and
scope of R’s resources being second to none.
6. R VISUALIZATION
Immediately visualize your data with functions
R was designed specifically for data
tailored for any graphic you will need. Use
manipulation. That said, you can do grammar of graphics alongside the ggplot2
the majority of operations you will package to create meaningful, highly-
customizable visualizations, and plots.
need in data science – like filtering,
Think outside the common static canvas, and
arranging, factoring, gathering,
design interactive graphs for the web – R does
mutating, spreading your data, and this fast and effectively, too.
much more – in a handful of lines of
code.

BIG DATA
R is designed to handle extremely
big data sets, usually gathered by

COMPUTATIONAL medium to large companies, or for

ANALYSIS academic research.

Enjoy the full computational power of your

computer, exponentially increasing the
speed of the analysis through specialized
libraries and vectorized code.
7. Python

Python is an open-source, general-

purpose high-level programming
language. It is one of the most widely
used programming languages in the
past few years. The technical
advantages it has over other
programming languages and its
modules for scientific computing make
it a preferred choice while working in
the fields of in finance, econometrics,
economics, data science and machine
learning.
7. Python Leading BI software Tableau launched TabPy which is an integration of
BI Python within Tableau. In this way, you can create reports and
dashboards that leverage the real-time computational power of Python.

Python is compatible with MySQL and is expected

to be integrated in the new version of the Microsoft
DBMS SQL Server as well. Thus, giving you the capability of
working with relational databases. Furthermore,

Python Libraries Python can be used to produce non-relational

Easy syntax, open-source,

databases, such as NoSQL.
integration, scope
NumPy, pandas, matplotlib, Seaborn,
scikit-learn, and TensorFlow are some Python is the leading language used for
of the most widely used libraries in machine learning in the field of data science.
data science. They combine the Machine With the amazing scikit-learn and powerful
capabilities of many other
learning frameworks like TensorFlow, Keras, and
programming languages, but let you PyTorch, Python is easily the all-in one
use them all in one place in an programming language for ML.
environment that just needs to
support Python.
8. Advanced
Statistics
Advanced statistics in this framework
refers to the symbiosis between linear
algebra, computational power, and
predictive modeling. Examples are
linear regression, logistic regression,
and clustering. Depending on the
source, this part of data science can be
called anything from statistical methods
to machine learning. Either way, the
techniques involved are the same and
are pretty much what professionals
understand by the term ‘data science’.
 Math
8. Advanced Statistics methods Distributions

Bayesian
Transform inference
Basic statistics lays the foundation of the field and
focuses on frequentist inference. Advanced
statistics builds up on it, entering
multidimensional spaces, through knowledge of Advanced
mathematical methods, transformations and
Regression
topics
distributions. Moreover, more complex means of Decision
theory
analysis are introduced, such as regression,
classification, clustering and factoring. Finally,
Bayesian inference and decision theory allow the
Data Scientist to solve problems of dynamic
and/or behavioral nature. Factor Classification
Cluster analysis
analysis
9. Machine learning
Machine learning is often confused with
artificial intelligence. In reality, machine
learning is a revolutionary approach to
developing AI programs, but is not the AI
itself. One of the definitions of machine
learning is: ‘extracting knowledge from
data”. In fact machine learning is closely
related to data mining and statistics. In the
context of data science, the Data Scientist
will be looking for ways to analyze the data
using machine learning algorithms, in order
to solve problems that are too complex or
incomprehensibly big for the human brain
to process.
9. Machine Learning
Machine learning is a relatively new field
that is constantly evolving. In order
Supervised Unsupervised Reinforcement*
create and run machine learning
algorithms, one needs solid statistical
In supervised ML, the In unsupervised ML, the In reinforcement ML, the
knowledge and programming skills. In algorithm’s goal is to find algorithm’s goal is to goal of the algorithm is to
the best way to perform reach a result, which is maximize its reward. It is
the field of data science, most often, the task given by the unknown to the inspired by human
researcher. It ‘learns’ what researcher. Once an behavior and the way
machine learning is divided into three
the approach is output is given, the data people change their
subsets: supervised, unsupervised, and (mathematically, finds the scientist is expected to actions according to
perfect fitting function for interpret what the incentives, such as getting
reinforcement machine learning. Each of the problem). program has done. a reward or avoiding
punishment.
them is based on different traditional Common methods: Common methods:
• Regression • Clustering Common methods:
statistical methods, thus has different
• Classification • Decision process
strong sides and shortcomings. • Reward system

*The literature on the topic divides machine learning into supervised and unsupervised. In AI frameworks, reinforcement is typically considered a subset of supervised and/or unsupervised. However, in the
field of data science, it is common to divide it in a distinct subset due to the nature of the methods used. That is also the classification that we have adopted.
10. SQL + Tableau
Knowledge of SQL and Tableau are two
indispensable skills for a Data Scientist.
What truly distinguishes an analyst from
his/her peers is interdisciplinary
knowledge, breadth, and ability to
combine expertise from different
domains. One of the most impressive
ways you can differentiate yourself from
other analysts is the ability to work with
data from the very source and then
present it through beautiful, meaningful
and professional visualizations.
10. SQL + Tableau

Plan Connect with SQL Preprocess Visualize

• Define the problem • Access the database in • Design KPIs • Plot the data
• Figure out how to Tableau • Preprocess the data in • Visualize professionally
acquire the data • Find all relevant calculated fields • Create dashboards
• Think about information
visualization
SQL

Plan your data journey. Once you connect your database Raw data is rarely suitable for It is up to you to create the
Professional visualization implies with Tableau, you can write queries visualization right away. More often visualizations intended in the beginning.
thinking your problem through to extract information directly in the you must use some degree of Tableau provides a seamless experience
from data collection to the axes interface. Additionally, you can preprocessing to design the metrics both for single plots and professional
of your final plots. combine several databases. you’ll later visualize. dashboards.
11. SQL + Tableau +
Python
Knowledge of SQL, Tableau, and Python
are all indispensable skills for a data
scientist. While traditionally different
activities are performed with each tool,
Tableau developers have worked hard on
integrating them. This advancement has
brought about new ways to present data.
Interactive dashboards and real-time
maps, created by machine learning
algorithms, coded in Python, with SQL
data, all visualized in Tableau. The
ultimate data science experience!
11. SQL + Tableau + Python The dashboard user inputs
data and gets a prediction

Similar to the integration of SQL and Tableau, data

scientists further add Python to the picture. Apart
Customer retention
from using Python in Tableau’s calculated fields, one
This is data from an Audiobook app.
can actually perform machine learning inside its Logically, it relates only to the audio
versions of books. Each customer in the
interface, or design models in Python and later database has made a purchase at least
once, that's why he/she is in the database.
visualize them and access them from Tableau. A deep learning algorithm based on our
available data predicts with 94% accuracy,
if a customer will buy again from the
This is an amazing presentation and storytelling tool Audiobook company.

in the data scientist arsenal. In the image on the right You can input Minutes listened, Price paid,
Review, Completion rate, and Support
you can see a beautiful dashboard with text, Tableau requests for a given customer.

graphs, and a customizable field, for users to input Reviews Sales

data and receive the answer provided by the

machine learning algorithm behind the scenes.

This is truly a bridge between data science and non-

technical audiences like no other!
FAQ at interviews

1. What does data science mean?

2. What are the assumptions of a linear regression?
3. What is the difference between factor analysis and cluster analysis?
4. What is an iterator generator?
5. Write down an SQL script to return data from two tables.
6. Draw graphs relevant to pay-per-click adverts and ticket purchases.
7. How would you explain Random Forest to a non-technical person
8. How can you prove that an improvement you introduced to a
model is actually working?
9. What is root cause analysis?
10. What is a logistic regression?
FAQ at interviews

11. Explain K-means.

12. What kind of RDBMS software do you have experience with?
What about non-relational databases?
13. Supervised learning vs unsupervised learning.
14. What is overfitting and how to fix it?
15. What is the difference between SQL, MySQL and SQL Server?
16. How would you start cleaning a big dataset?
17. Give examples where a false negative is more important than a
false positive, and vice versa.
18. State some biases that you are likely to encounter when
cleaning a database.
Are you ready to take the first step to your
future career today?