BUILD A PORTFOLIO

Learn what makes a good

data science project and get
project Ideas

IDENTIFY WHAT
TYPE OF LEARNING
STYLE FITS YOUR
PERSONALITY

And obtain a curated list of

resources that fits your
learning style

PREPARE FOR
INTERVIEWS

Get resume and interview

preparation tips
TABLE OF
CONTENTS

INTRODUCTION
1 About this guide 4
About me 7
Guide vs Roadmap 10
How this guide is organised 11

ROADMAP
2 What do data scientist do? 13
Difference between data scientist, data analyst, research
analyst, data engineer, and ML engineer 14
Skills required 16
R vs Python 20
Generalists vs Specialists 22
Known your domain 24

RESOURCES
3 Understanding your way of learning 28
Learning how to learn 31
Learning strategies 32
Growth mindset 35
Data science resources 37
TABLE OF
CONTENTS

STUDYING
4 Degree vs bootcamp vs self-taught 50
Do I need a PhD to get a data science job? 54
How long to learn data science from zero to job-ready? 55
Learning data science while working full-time 56
How to get out of (or completely avoid) the tutorial trap 59
Productivity tips 62
Common mistakes data scientists make and how to avoid
them 63

PORTFOLIO
5 What to include? 67
Where to publish your portfolio? 69
Project ideas 72
How to manage you portfolio projects 78

FIND A JOB
6 Job readiness checklist 82
How to find a a job 83
Prepare for interviews 89
Interview checklist 94
Salary negotiations 95
Imposter syndrome 98
Recognizing and avoiding burnout 99
Build Your Own Data Science Journey

CHAPTER I: INTRODUCTION

ABOUT THIS
GUIDE
We're living in a digital world. The global economy relies heavily on digital
technologies, and the amount of data we have is mind-blowing. It's already
2023, and it feels like we're living in the future. One of the most sought-after
jobs in the market right now is that of a data scientist. The demand for data
science professionals is massive and only expected to grow further. In fact,
it seems like the demand will outpace the number of available data
scientists. So, if you're thinking about a career change and becoming a data
scientist, now is the perfect time. This guide will walk you through the entire
process.

Drawing from my experience working as a data scientist and data analyst

with multiple companies, I've had the opportunity to witness the hiring
process for data scientists and the formation of data science teams. I know
what it takes to land your first job, the essential skills you should
acquire, what to prioritize, and what you should showcase during a
job interview. In this guide, I'll share this knowledge with you so that you
can also step into the fascinating world of data science - a market that holds
great promise for the future.

Page 04
About This Guide

If you're looking for a quick summary of how to become a data scientist,

here are the most important three steps:

1. Build your GitHub repository and start creating projects. You can
choose a dataset from Kaggle and work on building something around
it. Generally, classification problems tend to be more manageable. This
will allow you to refine your skills and demonstrate your commitment to
potential employers.
2. Engage in LinkedIn groups and Slack communities dedicated to data
science and machine learning. Look for online meetups and
conferences, or even attend local ones if possible to expand your
network and meet like-minded individuals. Networking is key to finding
job opportunities.
3. Code, code, and code some more! Data science is ultimately a practical
skill. Share your work on social media, Medium, GitHub, and LinkedIn to
increase your chances of finding a job.

From a technical standpoint, to start, you'll need to familiarize yourself with

the following:

Fundamentals of statistics and linear algebra

Data visualization tools like plotly, ggplot, or [Link]
Data processing and storage algorithms
Regression and classification algorithms in machine learning
Dimensionality reduction techniques like Principal Component Analysis
Basics of neural networks

I'll dive deeper into the specifics in the following chapters.

Page 05
About This Guide

Focus

This guide is specifically designed to answer questions on how to acquire

the necessary skills for data science and secure a job in the field. As the
emphasis is on career development, it intentionally steers away from the in-
depth technical nuances of data science. Thus, you won't find discussions
on model hyperparameters or detailed explanations of Python packages. In
fact, this guide does not include a single equation or line of code. But in this
guide I've recommended numerous great books and courses that delve into
these technical details. The intention of this guide is to shed light on the
often overlooked but equally important non-technical knowledge required
for success in data science. Additionally, this guide provides a clear
roadmap with step-by-step instructions to help you learn all the necessary
technical topics.

Who is this guide for?

The Build Your Own Data Science Journey guide is designed to assist
individuals who are interested in entering the field of data science and want
to advance their careers in it. Whether you're considering a career in data
science or have already spent a few years in the field, this guide is for you.

If you're an aspiring data scientist, this guide will provide you with the
necessary skills and knowledge to become a successful data scientist. And if
you're a junior data scientist looking to progress to a more senior role, this
guide will help you navigate the path to advancement.

By the time you finish reading this guide, you'll feel confident about the
steps you need to take to excel in your data science career. So, let's get
started and dive right in!

Page 06
SELF-TAUGHT
DATA
SCIENTIST

About me

Hey there! I'm Polina. I'm a self- I've created this guide to share the
taught data scientist, and you might strategies, tips, and resources I wish
know me from Instagram, where I I had known earlier in my career. My
share my data science journey. I goal is to answer your questions
started just like you, trying to and offer guidance as you embark
navigate this vast world of data on your own data science journey. I
science. This world seemed want to show you how to effectively
overwhelming at first, but I gained self-study and acquire the
valuable knowledge with each step I necessary skills.
took and each mistake I made.
About me

By the way, I didn't have a tech My area of expertise is causal

degree or attend a bootcamp. Yet, machine learning, and I am
after just two months of learning to fortunate enough to work on
code, I secured my first coding diverse projects ranging from
internship. Over the next two years, analysing vast quantities of data to
while continuously learning, I held employing the latest ML techniques.
various data roles and eventually
landed a full-time job as a data During my self-learning journey, I've
scientist in an established discovered efficient ways to learn
consulting company. quickly, what areas to focus on, and
how to avoid getting stuck in
endless tutorials. I've packed all of
these insights and more into this
guide. Following the steps outlined
in this guide will make you more
confident in programming and well-
prepared for your first data science
job. I am so excited about your
future success, let's get started!

Page 08
Start a data science career

Start a blog

Code, code, and code some more

Select learning resources

Attend conferences, meetups

Guide vs. Roadmap

The "Data Science Express" The Roadmap offers a structured

Roadmap and the Data Science map of topics to learn on your
Guide "Build Your Own Data journey to becoming a data
Science Journey" are separate scientist. It's designed to give you a
products; though they complement clear understanding of the
each other very well. You can concepts, skills, and areas of
download the roadmap on my knowledge you may want to cover.
website if you have not already.
On the other hand, the guide is a
comprehensive plan that goes
beyond just listing the topics. It
provides detailed instructions on
how to actually break into the field
of data science. This includes advice
on how to build a compelling
portfolio, what resources to use for
learning, what projects to
undertake, and more.

In essence, while the Roadmap

provides you with the 'what' (i.e.,
what topics to learn), the Guide
provides you with the 'how' (i.e.,
how to learn these topics and how
to use this knowledge to transition
into a data science career). While
they can be used separately, using
both the Roadmap and the Guide
together will give you a holistic,
structured, and practical approach
to entering the field of data science.

Page 10
 HOW THIS GUIDE
IS ORGANIZED
Chapter I: INTRODUCTION
This chapter provides background information about the author, the
guide itself, and what readers can expect from it.
Chapter II: ROADMAP
In this chapter, different paths within the data science domain are
explored, and the necessary skills to become a data scientist are
listed.
Chapter III: RESOURCES
Dedicated to learning styles and resources, this chapter includes a
Notion template (duplicated as a table within the guide) that readers
can use to create a personalized study curriculum.
Chapter IV: STUDYING
This chapter delves deep into how to study effectively, the strategies
to adopt, how to avoid falling into tutorial traps, and guidance on
choosing between a degree, bootcamp, and self-taught route.
Chapter V: PORTFOLIO
Focused on building a strong portfolio in data science, this chapter
provides insights on how to get started, stay motivated, which
projects to undertake, and how to stand out.
Chapter VI: FIND A JOB
Starting with resume building and ending with nailing the interview,
this chapter guides readers through the complex job search
process in the field of data science.
CLOSING REMARKS

Page 11
 II

The roadmap is more important

than the road.

Chapter II
ROADMAP
Build Your Own Data Science Journey

CHAPTER II: ROADMAP

WHAT DO
DATA SCIENTIST DO
A data scientist is an expert with a deep knowledge of data, algorithms, and
data visualization. They possess the ability to work as part of a team,
understand data structures, analyze data, design and create charts and
graphs, and write concise code.

There are broadly two types of data scientists:

Scientific Researchers: These data scientists combine statistical

methods and software development expertise to create, analyze, and
interpret complex datasets in order to advance knowledge in a specific
field of study.

Industry Data Scientists: These data scientists are professional who

leverages statistical, machine learning, and data analysis techniques to
extract valuable insights from real-world data to address business-
related challenges and opportunities.

Both types are great; they just require a little bit of a different skillset. I am
more of the first type. The type of data scientist you become will depend on
where you want to work and in what team. For the first type, it is important
to focus on math and stats skills. For the second type, focus on MLOps, data
pipelines, and visualization skills.

Page 13
Difference between Data Scientist, Data
Analyst, Research Analyst, Data Engineer,
and ML Engineer:

1. Data Scientist:
Role: Uses advanced computational and statistical methods to find
patterns and insights in data.
Skills: Typically has expertise in machine learning, math, statistics, and
programming.
Tasks: Builds models to predict future events, derives insights from
complex and unstructured data, often uses machine learning and deep
learning frameworks.
2. Data Analyst:
Role: Interprets data to answer specific questions, usually involving
structured data.
Skills: Proficient in statistics, data visualization tools, and basic
programming.
Tasks: Creates reports using charts and graphs to visualize data,
conducts A/B testing, uses descriptive statistics to answer business
questions.
3. Research Analyst:
Role: Investigates specific industry or market conditions to help
organizations make informed decisions based on their research.
Skills: Strong in research methodologies, domain-specific knowledge,
data interpretation, and often, report writing.
Tasks: Gathers qualitative and quantitative data about a subject,
provides insights and recommendations based on findings, often works
in finance, marketing, or other specific sectors.

Page 14
4. Data Engineer:
Role: Focuses on the practical application of data collection and data
processing.
Skills: Expertise in databases (SQL, NoSQL), data warehousing
solutions, and big data technologies like Hadoop and Spark.
Tasks: Designs, constructs, installs, and maintains large-scale
processing systems and other infrastructure; ensures data is clean,
reliable, and easily accessible.
5. ML Engineer (Machine Learning Engineer):
Role: Bridges the gap between data science and software engineering.
Focuses on designing, building, and deploying machine learning models
at scale.
Skills: Proficient in machine learning, software development, and cloud
solutions.
Tasks: Takes data science prototypes and turns them into scalable
production applications, optimizes ML algorithms, works closely with
data scientists to understand and implement algorithms.

In essence, while there's some overlap among these roles, each has its
distinct focus and set of responsibilities. A data scientist and ML engineer
might both work with machine learning models, but the former emphasizes
model development and the latter on deployment. A data analyst is closer
to the business side with descriptive statistics, while a data engineer
ensures that data is properly stored, processed, and made accessible. A
research analyst, on the other hand, is more domain-specific, delving into
market or industry-specific data to draw conclusions.

Page 15
Skills Required

If you want a list of topics you If you can use Python or R, know
should learn, refer to my Data some SQL, and understand the
Science Express roadmap, which basics of ML algorithms, you're on
you can download here for free. the right track.
But broadly speaking, the required
skills largely depend on the specific If you can plan your career by
job role and the employer's learning more things up front, that
expectations. Top-tier companies, in is always helpful, but I do not think
general, have high standards. They it is a requirement. What truly
might be looking for candidates matters is your dedication to
proficient in Python, R, ETL continuous learning. If you are
pipelines, data engineering, trying to get hired by top tech
experiment design, causal companies, you need a little more,
inference, and capable of designing but that is more for your resume to
and deploying models. It's quite a stand out. It is helpful to distinguish
list! between the core skills you need to
start your career in data science
But to start your career in data and others that are nice to have if
science, you don't need to know you want to get into a competitive
everything. brand-name company.
Crafting a Balanced Data Science Skillset
How do you make sure that you have a well-rounded skillset? Fill out the
"Wheel of Data Science" provided on the next pages!

The "Wheel of Life" is typically a holistic tool used for self-assessment in

various aspects of one's life. For a data science context, we can adapt it into
a "Wheel of Data Science" to focus on the various aspects that contribute to
a well-rounded skillset and career in this field. Here's how you can break it
down:

1. Programming Skills
How proficient are you in programming languages commonly used in data
science, such as Python or R?
2. Statistics
How well do you understand the statistics that underpin the analyses you'll
be performing?
3. Machine Learning
Rate your expertise in machine learning. Can you create models, select
algorithms, and understand their pros and cons?
4. Communication
Can you explain technical results in a way that non-technical people can
understand?
5. Domain Expertise
Do you have sufficient knowledge in the industry you are working in or
aiming to work in? (e.g., healthcare, finance, retail)
6. Dataviz
How well can you create meaningful graphs, dashboards, or other types of
data visualizations?
7. Project Management
How good are you at planning, executing, and completing data science
projects?
8. Math
Are you comfortable with mathematical concepts, like linear algebra,
calculus, and discrete mathematics, which form the foundation for many
data science algorithms and techniques?

Page 17
Crafting a Balanced Data Science Skillset
How to Use the Wheel
Draw a circle and divide it into 8 equal parts, one for each category.
Rate Yourself: On a scale of 1 to 10, fill in each section with how satisfied
or proficient you are in each aspect. The center of the circle is a 1, and
the outer edge is a 10.
Reflect: Look at the filled-in wheel. Is it balanced? Are there areas that
need more attention than others?
Set Goals: For the sections where you rated yourself lower, set
achievable goals to improve those areas.
Review and Adjust: Regularly come back to your wheel to track your
progress and make adjustments as needed.
By systematically assessing and setting goals for each aspect of your data
science learning journey, you'll be better positioned for well-rounded
growth and success in the field.

Page 18
 WHEEL OF

DATA SCIENCE
THE WHEEL OF DATA SCIENCE IS A GREAT TOOL THAT HELPS YOU
BETTER UNDERSTAND WHAT YOU CAN DO TO MAKE YOUR SKILLS MORE
BALANCED. THINK ABOUT THE 8 SKILLS CATEGORIES BELOW, AND
RATE THEM FROM 1 - 10.

SK ILLS STA
M ING TIS
M TIC
A S
GR
RO
P

M
AC
HI
TH

NE
MA

LEA
RNING
1 2 3 4 5 6 7 8 9 10
MENT

COMM
GE
NA

N U
MA

I C A
T

TI
C
JE

N O
RO

P DO
MA
Z I IN
AV EX
DAT TISE
PER
R vs Python

The age-old debate among data Moreover, there's a general

scientists often revolves around a consensus that transitioning from
simple question: Should I learn R or Python to R is more straightforward
Python? As you plunge into the vast than the other way around. If you
ocean of data science, this query is begin with Python, you're setting up
a natural starting point. Here's a a solid foundation, making it easier
short and sweet answer to ease to pick up R later if needed.
your dilemma: Python. However,
the full story offers more nuance. That being said, R is far from
obsolete. It has its own strengths
Firstly, Python is more in demand. A that make it appealing, especially
quick glance at job postings or for those delving deeper into
discussions in the data science statistical analysis.
community confirms Python's R has a longstanding history in
dominance in the industry. Its the realm of statistics and
versatile nature allows it to be research. Because of this, many
employed not just in data analytics statistical models and
but also in web development, techniques have already been
automation, and several other encapsulated into R packages.
domains. This broad applicability This wealth of packages can
enhances its popularity. sometimes give R an edge,
especially for highly specialized
analyses.

Page 20
R vs Python

R is often the language of Both languages have their place,

choice in academia. Cutting- and proficiency in each can be a
edge methodologies and strong asset. Python, given its
breakthroughs emerging from versatile nature, tends to be more
the research community adept at handling large datasets. Its
typically come with R robust libraries and frameworks
implementations. If you're also make deploying models more
inclined towards research or seamless, catering to real-world
aim to be at the forefront of applications of data science. On the
data science innovation, other hand, R's deep roots in the
familiarity with R can be statistical community mean that it
beneficial. can sometimes offer a more native
approach to specific analytical tasks.
While Python's dominance is clear,
it's essential to recognize that some To kickstart your data science
companies are language agnostic. journey, Python is a recommended
They value your ability to extract starting point given its demand and
insights from data, develop models, broader applications. However, as
and deploy solutions more than you progress, gaining proficiency in
your language preference. For R can open doors to specialized
instance, in my own experience, I've analysis.
worked on projects requiring both
Python and R.

Page 21
 Generalists vs.
Specialists in
Data Science
So you're diving into the world of data science
and you've hit a crossroads: should you aim to
be a generalist who knows a little bit of
everything, or should you be a specialist who
really digs deep into one area?

What's the Difference?

Let's keep it simple. A generalist is someone who can handle different data
tasks from cleaning up data, to creating visualizations, to building machine
learning models, to even designing data pipelines. You are a data scientist,
software developer, and data engineer at the same time kind of. You're a
valuable team member when there's a lot to do and variety is key.

Specialists, meanwhile, excel in a specific area. If you specialize in something

like Causal Machine Learning, you're the person everyone goes to with that
particular question about treatment effects or causality. There are other paths
to specialize in too, such as Natural Language Processing (like building
chatbots), Deep Learning (focused on things like image recognition), or Time-
Series Analysis (analyzing trends over time).

Page 22
Generalists vs. Specialists in Data Science

Pros and Cons. Being a generalist has its benefits. You can fit into different
roles and contribute to various kinds of projects. If you're part of a small
team or a startup, your diverse skills will come in handy. On the downside,
you might not be the go-to expert when a complex, specialized problem
arises.
On the flip side, specialists often become the experts in specific problems
and their skills are usually in high demand. They can ask for higher salaries
for their expertise. However, they also risk their skills becoming less
relevant if the industry shifts its focus.

What Should You Consider? Your career stage matters. If you're just
starting out, being a generalist can give you a broad perspective and help
you figure out what you really enjoy. As you gain more experience, you may
feel drawn to a specialized area, and that can be a good time to dive deep.
For those later in their careers, being a specialist can also open doors to
consulting roles or positions that require deep expertise.
Keep an eye on the job market too. What skills are in demand? Additionally,
consider what you enjoy—do you like tackling different kinds of tasks or do
you prefer to go deep into a subject?

Climbing the Ladder. Here's the important question— who's more likely
to land senior roles, generalists or specialists? The answer varies. In larger
organizations where roles are well-defined, specialists often have a clearer
path to senior positions. However, in smaller organizations, the broad skill
set of a generalist can be invaluable for leadership roles.

Choosing between being a generalist and a specialist isn't a one-time

decision. Your career is a journey, and it's totally fine to shift your focus over
time. Some people even find a middle ground, being generalists with deep
expertise in a couple of areas. The most important thing is to keep learning
and adapting because the world of data science is always changing.

Page 23
Know Your Domain

Whether you decide to be a Data never exists in a vacuum.

generalist or a specialist there's one When you're looking at numbers or
thing both paths have in common: trends, it's so easy to forget that
the need to know your domain. behind every data point is a real-
Understanding the industry or world situation. If you're working in
sector you're working in is crucial retail, understanding consumer
for any data scientist. behavior and market trends will
make your data analysis more
Imagine you're a data scientist in meaningful. In the medical field,
healthcare, but you don't know the knowing the specifics about various
difference between outpatient and diseases or treatments can be
inpatient care. Or picture working in crucial when interpreting healthcare
finance but not understanding what data.
ROI means. You might have all the
data science skills in the world, but One of the main goals of data
without understanding the specifics science is to help make better
of your domain, your insights could decisions. Domain knowledge is
miss the mark. Knowing your what bridges the gap between raw
domain helps you ask the right data and useful insights.
questions, understand the nuances
in your data, and ultimately, offer
more valuable solutions.

Page 24
Know Your Domain

For example, if you're working in If you're at the beginning stages of

finance, understanding economic your career, choosing projects that
indicators or market volatility can allow you to work closely with
help you build more accurate domain experts can be invaluable.
predictive models. You're not just For those further along, perhaps
crunching numbers; you're consider taking some courses or
interpreting them in a way that's earning certifications that are
valuable to your industry. specific to your domain. Some data
scientists even have dual degrees,
Now, the million-dollar question— like one in data science and another
how do you actually gain this in healthcare or finance, to bolster
domain knowledge? Well, the first their domain expertise.
step is to start learning about the
industry you're in or interested in. If you're making a career switch into
Read up on trends, key terms, and data science, don't think of it as
major players. Don’t underestimate starting from scratch. In fact, you
the value of talking to experts in the might be luckier than you think.
field either. They can provide Your past experience in another
insights you won’t find in any report domain can actually be a major
or article. asset.

Page 25
Know Your Domain

Let’s say you worked in government You're not just another data
before making the jump to data scientist; you're a data scientist who
science. You already have a strong understands the intricacies of
understanding of laws, regulations, governmental processes and
and policies. Now, combine that policies. You can offer not just
with your new data science skills, technical solutions but meaningful
and you become an extremely insights tailored to the specifics of
valuable asset in the realm of policy the policy landscape. Your career
consulting. Imagine the possibilities: switch isn’t a reset; it's a level-up,
you could be crunching numbers to blending your past experience with
help policymakers and stakeholders new skills to make you more
determine the viability of a new bill versatile and insightful than ever.
or the effectiveness of an existing
program. Your domain knowledge So what if you're new to data
could help you pose relevant science or perhaps you're making a
questions that others might career switch and don't have a
overlook, or foresee implications domain that naturally carries over?
that are not immediately obvious Don't worry, you're not alone, and
from the data alone. This is where this is actually an exciting
your expertise becomes a unique opportunity to shape your career
selling point. path deliberately.

Page 26
Know Your Domain

Choosing a domain might seem like One of the best ways to understand
a huge decision, and it is, but it's a domain is to talk to people who
also one that you can make are already working in it. Don't
thoughtfully. hesitate to reach out to
professionals in the field for
The first step is to think about what informational interviews.
genuinely interests you. Are you
fascinated by human behavior? Before you fully commit, see if you
Maybe something like consumer can work on some projects or take
analytics or psychology-focused some courses related to the
research could be a fit. Do you get domain you're considering. Many
excited when you think about online platforms offer domain-
technological advances and their specific data science courses that
potential impact? Then perhaps a can give you a taste of what working
domain like tech or AI ethics would in that field might be like.
be up your alley.
Choosing a domain is not a lifetime
While following your passion is sentence; many data scientists
important, you also have to switch domains throughout their
consider the practical aspect. Look careers. However, making a
at what skills and domains are in thoughtful choice can give you a
demand. Sites like LinkedIn or job strong start and make your skill set
boards can give you an idea of what particularly attractive to employers.
employers are looking for. Whether you find your domain
through your past career, personal
Another important consideration is interests, or even a bit of trial and
the future of the domain. Is it a error, it will play an integral role in
growing field with a lot of shaping your path in data science.
opportunities? You don't want to
invest in learning a domain that will
become obsolete in a few years.

Page 27
Build Your Own Data Science Journey

CHAPTER III: RESOURCES

Understanding
Your Way of
Learning
Learning data science can often seem tricky and confusing because there's
a lot to cover—from statistics, coding, and machine learning to creating
data visualisations, and more. This chapter is here to make this learning
journey easier. We'll talk about why it's important to know how you learn
best before you choose resources and strategies to learn data science.

Each of us has our own way of sorting, and remembering information.

There's no one right way to learn, which is why it's crucial to understand
how you learn best before you start your data science journey. Knowing
your learning style lets you tailor your learning experience, making it more
efficient, effective, and even fun. Knowing how you learn best won't
instantly make learning data science easy, but it will give you the tools to
approach your learning in a more targeted and efficient way. Keep in mind
that the goal isn't to learn as fast as you can, but to really understand and
apply what you learn.

In the following pages, you'll find a quiz designed to identify your best-suited
learning style. Understanding this will guide you in selecting the most
suitable resources, enhancing your overall learning experience in data
science.
Page 28
 IDENTIFY YOUR LEARNING STYLE

QUIZ
Single Choice Questions

WHEN TRYING TO UNDERSTAND NEW INFORMATION, I:

A. Prefer to read or write about it. C. Like to discuss it and talk it through.
B. Need to see charts or diagrams. D. Have to get hands-on and try it out.

IN A CLASSROOM OR MEETING, I:

A. Enjoy taking detailed notes. C. Benefit most from group discussions.

B. Focus on the slides or visual aids. D. Prefer participating in interactive
activities or experiments.

I REMEMBER INFORMATION BEST WHEN I:

A. Write it down or read it C. Explain it to others.

several times. D. Apply it or do it in real life.
B. Visualize it or create a mental
image.

WHEN STUDYING OR PREPARING FOR A TASK, I:

A. Write summaries or rewrite notes. C. Discuss topics with others or

B. Draw mind maps or diagrams. teach someone.
D. Use flashcards or practical exercises.

Page 29
 A FEW MORE QUESTIONS

IN MY FREE TIME, I PREFER TO:

A. Read a book or write in a journal. C. Talk to friends or listen to music.

B. Watch movies or play video D. Engage in a hands-on hobby or
games. play a sport.

WHEN TRYING TO RECALL A PAST EVENT, I:

A. Remember the words C. Recall the discussions or debates.

or conversations. D. Remember what I was doing.
B. Visualize the surroundings and
what was happening.

WHEN GIVING DIRECTIONS, I:

A. Write them down. C. Verbally explain them.

B. Draw a map. D. Demonstrate or walk them through
it.

Results:
Mostly A's: Read/Write Learner - You prefer to learn through words. Reading and
writing help solidify concepts for you.
Mostly B's: Visual Learner - Images, diagrams, and visual cues are your go-to. You
understand concepts better when you see them.
Mostly C's: Auditory Learner - You learn best by listening. Lectures, discussions, and
oral explanations resonate with you.
Mostly D's: Kinesthetic Learner - You prefer to be actively involved in the learning
process. Experiencing things firsthand is key for you.

Remember, many people have a mix of learning styles, and this quiz is a basic overview.
It's always beneficial to use diverse learning strategies to enhance understanding and
retention.

Page 30
Learning How to Learn

Learning is not a one-size-fits-all Determining your learning style is

process. It's a unique journey that not about boxing yourself into a
varies from person to person. We category, but rather understanding
all have different ways of which methods help you learn most
understanding and retaining new effectively. The quiz above can help
information. Identifying these you identify your dominant learning
personal learning strategies and style.
styles can make the learning
process more enjoyable and Beyond learning styles, it's also
efficient. important to consider your learning
strategies. These are the methods
There are four primary learning you use to learn new information.
styles: visual, auditory, read/write, For example, some people prefer to
and kinesthetic. Visual learners immerse themselves in one topic at
prefer using images and diagrams, a time (blocked learning), while
auditory learners thrive on listening others might prefer to learn
and speaking, read/write learners multiple topics simultaneously
favor reading and note-taking, and (interleaved learning).
kinesthetic learners need a hands- Understanding your preferred
on approach, learning through learning strategies can help you
experience and practice. organize your study schedule and
material in a way that enhances
your learning.

Page 31
Learning Strategies

Once you've identified your learning But being conscious of these

style and strategies, you can tailor strategies and expanding your
your learning experience toolkit can greatly enhance your
accordingly. For instance, if you're a ability to learn, especially when
visual learner who prefers blocked tackling a complex field like data
learning, you might decide to focus science.
on one topic at a time, using lots of
diagrams and visual aids. Deep Learning: Deep learning
Alternatively, if you're an auditory involves actively seeking to
learner who prefers interleaved understand the meaning or
learning, you might choose to study underlying principles behind the
various topics in parallel, using information. Deep learners often
resources like podcasts and draw connections between new
lectures. knowledge and what they already
know, making it more likely for them
Learning strategies are systematic to recall and apply this information
approaches we use to understand, later. Data Science requires a deep
remember, and apply new understanding of certain concepts,
information or skills. Each person especially in areas like statistics,
has a unique mix of learning machine learning, and algorithms.
strategies that they naturally
gravitate towards.

Page 32
Learning Strategies

It's important to go beyond merely you would learn a bit about linear
memorizing formulas or code regression, logistic regression, and
syntax. Try to understand the decision trees, for example, in the
underlying principles and why same study period. This method
things work the way they do. When encourages the brain to
learning a new coding concept, differentiate and identify the
always try to comprehend what similarities and differences among
each line of code does and why it's the algorithms, thereby improving
necessary. your understanding and retention
of each. Data Science is an
Interleaved and Blocked interdisciplinary field, combining
Learning: Blocked learning involves coding, statistics, machine learning,
focusing on one type of topic or skill and domain knowledge. Instead of
at a time before moving onto the mastering one area at a time,
next. If you're studying machine consider studying these areas in
learning algorithms, blocked parallel. For instance, as you're
learning would involve focusing learning Python, you can also start
deeply on one algorithm at a time. understanding basic statistical
concepts or machine learning
You might spend a week learning principles. This approach gives you
about linear regression, mastering a balanced skillset and allows you to
its concepts and its implementation, draw connections between different
before moving on to the next areas.
algorithm like logistic regression.
Interleaved learning, on the other Spaced Repetition: Spaced
hand, involves mixing different repetition involves revisiting and
types of topics or skills in your study reviewing information at increasing
sessions. In the same machine intervals over time. Instead of
learning context, interleaved cramming all your studying into one
learning would mean you study long session, it's usually more
multiple algorithms in each study effective to space out your study
session. Instead of mastering one sessions over a period of time.
before moving on to the next,

Page 33
Learning Strategies

This approach helps move Try to explain the concept in your

information from your short-term own words without looking at the
memory to your long-term memory. material. You can also use
Coding, like any language, requires flashcards with questions on one
regular practice for long-term side and answers on the other. This
retention. Spaced repetition can be method can be particularly helpful
especially useful here. For example, when preparing for data science
if you've just learned about loops in interviews.
Python, revisit this concept after a
day, then a week, and then a Elaborative Interrogation:
month. Coding platforms that offer Elaborative interrogation involves
daily coding problems can help generating explanations for why a
facilitate this approach. fact or concept is true. It
encourages you to link new
Active Recall: Active recall involves information with what you already
actively trying to remember know, creating a stronger memory
information from memory, without trace. Data science often involves
looking at the material. This might problem-solving, whether you're
involve testing yourself with debugging code or figuring out why
flashcards or trying to summarize a a model isn't performing well.
concept from memory. Active recall Elaborative interrogation can be a
has been found to be one of the powerful tool here. Always ask
most effective strategies for long- yourself why something is
term retention. When studying happening and try to come up with
theoretical concepts in data explanations. This encourages you
science, like different types of to think critically and problem-solve
machine learning algorithms, use more effectively.
active recall.

Remember, the journey to learning data science is not a sprint; it's a

marathon. It's not about learning everything quickly, but about
building a strong, lasting understanding of the key concepts. Knowing
your learning style and strategies is a vital first step towards this goal.

Page 34
 Growth
Mindset
Learning data science involves acquiring a
wide range of skills and tackling complex
problems, which can be extremely challenging.
However, remember that your own attitude
towards these challenges plays a significant
role in your success.

The term "growth mindset" was coined by psychologist Carol Dweck. It

represents the belief that intelligence and abilities can be developed through
dedication, hard work, and the right learning strategies. This is contrasted with
a "fixed mindset," the belief that intelligence and abilities are static and
unchangeable.

Data science is a rapidly evolving field that requires continuous learning. A

growth mindset is beneficial for several reasons:
Overcoming challenges: Learning data science is difficult. A growth
mindset helps you see challenges as opportunities to learn and grow rather
than as obstacles.
Dealing with failure: In data science, you'll encounter setbacks, such as a
model not performing as expected or struggling to grasp a new concept.
With a growth mindset, you can view these not as failures, but as part of the
learning process.
Keeping up with the field: The field of data science is always advancing. A
growth mindset encourages you to continuously learn and adapt, helping
you stay current with the latest developments.
Page 35
Developing a growth mindset involves changing how you think about
learning and challenges. Here are some strategies:
Embrace challenges: Instead of avoiding difficult tasks, embrace them.
See them as opportunities to learn and improve.
Learn from criticism: Seek out feedback and use it as a tool for
improvement. Try to view criticism as helpful information, not as a
personal attack.
Celebrate effort, not just results: Value the process of learning and
the effort you put into it, not just the end result. This helps keep you
motivated even when progress seems slow.
Persevere in the face of setbacks: When you face a setback, don't
give up. Instead, try to figure out what went wrong and how you can
overcome the obstacle.

A growth mindset is a powerful tool in your data science learning journey. It

allows you to embrace challenges, persist in the face of setbacks, and see
effort as a path to mastery. By cultivating a growth mindset, you'll be more
resilient, adaptable, and ultimately more successful in your quest to
become a data scientist. So remember, it's not just about learning data
science - it's about learning how to learn, and learning how to grow.
Page 36
 Data Science

Resourses

III

As you move forward,

remember to stay
flexible: be open to
trying new resources
and tweaking your
strategies to best fit
your changing needs
and goals.
The Depths of Data Science Mastery
Above the Water

Programming Basics
Basic Statistics
Data Visualization
Simple Machine Learning Models
Data Cleaning

Just Below the Waterline

Advanced Programming
Intermediate Statistics
Feature Engineering
Machine Learning
Data Wrangling

Deep Below the Surface

Deep Learning
Big Data Technologies
Model Optimization
Ethics in Data Science
Research & Development

At the Base

Mathematical Foundations
Domain Expertise
Critical Thinking
Continuous Learning
Collaboration & Communication
This Notion dashboard is a little gift from @polinacsv. Feel free to
use it as a template to craft a tailored data science curriculum!
Here is your link to the dashboard.
If Notion isn't your preference, you can find an identical
dashboard presented as a table on the subsequent pages. Begin
by picking resources that align with your primary and secondary
learning styles. From there, select between one and three
resources for every topic.
Have fun creating your personalized curriculum!

Page 39
 RESOURSES

Name Topic Type Level Cost

READ/WRITE LEARNERS

PYTHON

Designing Machine Learning

MLOps BOOKS Advanced PAID
Systems by Chip Huyen

TEXT-BASED
Made With ML MLOps Intermediate FREE
COURSES

Hands-On Machine Learning

with Scikit-Learn, Keras, and
READ/WRITE
ML LEARNERS
BOOKS Intermediate PAID
TensorFlow, 3rd Edition, by
Aurélien Géron

INTERACTIVE
Data Scientist in Python Career
ML CODING Intermediate PAID
Path by Dataquest
COURSES

Google's Machine Learning TEXT-BASED

ML Intermediate FREE
Courses COURSES

The Machine Learning

Simplified: A Gentle
ML BOOKS Introductory FREE
Introduction to Supervised
Learning by Andrew Wolf

Think
Bayes, 2nd Edition by Allen B. STATS BOOKS Advanced FREE
Downey

An Introduction to Statistical
Learning, with Applications in
Python by Gareth James,
STATS BOOKS Intermediate FREE
Daniela Witten, Trevor Hastie,
Rob Tibshirani, and Jonathan
Taylor

Essential Math for Data Science

MATH BOOKS Intermediate PAID
by Thomas Nield
 RESOURSES

Name Topic Type Level Cost

INTERACTIVE
College Algebra with Python MATH CODING Introductory FREE
COURSES

Interactive Dashboards and

Data Apps with Plotly and Dash DATAVIZ BOOKS Advanced PAID
by Elias Dabbas

TEXT-BASED
Kaggle Learn: Data Visualization DATAVIZ Introductory FREE
COURSES

Data INTERACTIVE
DATA
Analyst in Python Career Path CODING Introductory PAID
ANALYSIS
by Dataquest COURSES
READ/WRITE LEARNERS

Python
DATA
for Data Analysis, 3rd Edition, BOOKS Introductory FREE
ANALYSIS
by Wes McKinney

Fluent Python, 2nd Edition, by

CODING BOOKS Intermediate PAID
Luciano Ramalho

Python TEXT-BASED
CODING Introductory FREE
Tutorial by geeksforgeeks COURSES

TEXT-BASED
vetiver on CRAN MLOps Intermediate FREE
COURSES

Practical Statistics for Data

Scientists, 2nd Edition
ML BOOKS Introductory PAID
by Peter Bruce, Andrew Bruce,
Peter Gedeck

Hands-On Machine Learning

with R by Bradley Boehmke and ML BOOKS Intermediate FREE
Brandon Greenwell

Causal Inference:The Mixtape

STATS BOOKS Advanced FREE
by Scott Cunningham
 RESOURSES

Name Topic Type Level Cost

An
Introduction to Statistical
Learning, with Applications in R,
2nd Edition, by STATS BOOKS Intermediate FREE
Gareth James, Daniela Witten,
Trevor Hastie, and Rob
Tibshirani

An
Introduction to Bayesian
STATS BOOKS Intermediate FREE
Reasoning and Methods by
Kevin Ross

Introduction to R for
READ/WRITE LEARNERS
Econometrics MATH BOOKS Introductory FREE
by Kieran Marray

Fundamentals
DATAVIZ BOOKS Intermediate FREE
of Data Visualization by Wilke

INTERACTIVE
Data DATA
CODING Introductory PAID
Analyst in R Path on DataQuest ANALYSIS
COURSES

Advanced
R, 2nd Edition, by Hadley CODING BOOKS Advanced FREE
Wickham

R for
Data Science, 2nd Edition, by
Hadley Wickham, Mine CODING BOOKS Introductory FREE
Çetinkaya-Rundel, and
Garrett Grolemund
 RESOURSES

Name Topic Type Level Cost

VISUAL LEARNERS

PYTHON

ML Engineering for Production VIDEO

MLOps Advanced FREE
(MLOps) on Coursera COURSES

Harvard
VIDEO
University: Introduction to ML Introductory FREE
COURSES
Data Science with Python

Kaggle TEXT-BASED
READ/WRITE
ML LEARNERS Introductory FREE
Learn: ML courses COURSES

MIT:
Machine Learning with Python- VIDEO
ML Intermediate FREE
From Linear Models to Deep COURSES
Learning

MIT:
Data Analysis: Statistical VIDEO
STATS Intermediate FREE
Modeling and Computation in COURSES
Applications on EdX

Lecture
Notes for Causality in Machine STATS BOOKS Advanced FREE
Learning by Robert Ness

Mathematics
for Machine Learning VIDEO
MATH Intermediate FREE
Specialization by Imperial COURSES
College London on Coursera

From TEXT-BASED
DATAVIZ Intermediate FREE
Data to Viz: Python Gallery COURSES

Data
DATA VIDEO
Analysis on Introductory FREE
ANALYSIS COURSES
[Link]
 RESOURSES

Name Topic Type Level Cost

MIT:
Introduction to Computer VIDEO
CODING Introductory FREE
Science and Programming COURSES
Using Python on EdX

MLOps VIDEO
MLOps Intermediate FREE
with vetiver by Posit COURSES

Applied
Data Science with R VIDEO
ML Introductory FREE
Specialization by IBM on COURSES
READ/WRITE LEARNERS
Coursera

Lectures
VIDEO
on causality by Jonas Peters STATS Advanced FREE
COURSES
(MIT)

Introduction
to Econometrics with R by
Christoph Hanck, Martin Arnold, STATS BOOKS Introductory FREE
Alexander Gerber,
and Martin Schmelzer

TEXT-BASED
Ximera MATH Intermediate FREE
COURSES

From TEXT-BASED
DATAVIZ Intermediate FREE
Data to Viz: R Gallery COURSES

Data
DATA VIDEO
Analysis with R by IBM on Introductory FREE
ANALYSIS COURSES
Coursera

Data Science for Economists

and Other Animals CODING BOOKS Intermediate FREE
by Grant McDermott
 RESOURSES

Name Topic Type Level Cost

AUDITORY LEARNERS

PYTHON

Practical
VIDEO
Data Science on the AWS MLOps Advanced FREE
COURSES
Cloud Specialization

Preparing
for Google Cloud Certification:
VIDEO
Machine Learning Engineer ML Intermediate FREE
COURSES
Professional
READ/WRITE LEARNERS
Certificate on Coursera

Machine
VIDEO
Learning Specialization by ML Introductory FREE
COURSES
Andrew Ng on Coursera

Time
Series Forecasting in Python by STATS AUDIOBOOKS Advanced PAID
Marco Peixeiro

Khan VIDEO
MATH Introductory FREE
Academy COURSES

College
VIDEO
Algebra with Python on MATH Introductory FREE
COURSES
freecodecamp

Data
DATA VIDEO
Analysis with Python on Introductory FREE
ANALYSIS COURSES
freecodecamp

Practices
of the Python Pro, 1st Edition, CODING AUDIOBOOKS Advanced PAID
by Dane Hillard
 RESOURSES

Name Topic Type Level Cost

Web Apps
in R: Building your First Web
VIDEO
Application in R | Shiny Tutorial MLOps Intermediate FREE
COURSES
Series by
Data Professor

Mixtape ONLINE
STATS Advanced PAID
Sessions Webinars COURSES

Introduction
VIDEO
to Mathematical Thinking by MATH Introductory FREE
READ/WRITE LEARNERS
COURSES
Stanford on Coursera

Building
VIDEO
Data Visualization Tools by DATAVIZ Advanced FREE
COURSES
Johns Hopkins on Coursera

Data
Science: Foundations using R
DATA VIDEO
Specialization by Johns Hopkins Intermediate FREE
ANALYSIS COURSES
University on
Coursera

R
VIDEO
Programming by Johns CODING Introductory FREE
COURSES
Hopkins University on Coursera
 RESOURSES

Name Topic Type Level Cost

KINESTHETIC LEARNERS

PYTHON

INTERACTIVE
MLOps Fundamentals Track by
MLOps CODING Intermediate PAID
DataCamp
COURSES

Machine INTERACTIVE
learning in Python with scikit- ML CODING Intermediate FREE
learn by Inria COURSES
READ/WRITE LEARNERS
Modeling
and Simulation in Python by STATS BOOKS Intermediate FREE
Allen B. Downey

Think
Complexity 2nd Edition:
Exploring Complexity Science MATH BOOKS Advanced FREE
with Python by Allen B.
Downey

Data INTERACTIVE
Visualization with Python on DATAVIZ CODING Intermediate PAID
DataCamp COURSES

Data INTERACTIVE
DATA
Scientist Professional with CODING Advanced PAID
ANALYSIS
Python on DataCamp COURSES

INTERACTIVE
Python
CODING CODING Intermediate PAID
Programmer on DataCamp
COURSES
 RESOURSES

Name Topic Type Level Cost

INTERACTIVE
How to implement MLOps in R
MLOps CODING Advanced FREE
step by step
COURSES

Building INTERACTIVE
Web Applications with Shiny in MLOps CODING Intermediate PAID
R COURSES

MLU TEXT-BASED
ML Introductory FREE
explain COURSES

READ/WRITE LEARNERS
Statistical
TEXT-BASED
Methods in Psychological STATS Introductory FREE
COURSES
Research by Matthew Crump

Basic TEXT-BASED
STATS Introductory FREE
stats explained (in R) COURSES

TEXT-BASED
Mathigon MATH Introductory FREE
COURSES

INTERACTIVE
Data Visualization with R on
DATAVIZ CODING Introductory PAID
DataCamp
COURSES

INTERACTIVE
Data DATA
CODING Intermediate PAID
Analyst with R on DataCamp ANALYSIS
COURSES

INTERACTIVE
Posit Primers CODING CODING Intermediate FREE
COURSES
 IV

Focus on being productive instead

of busy.

Chapter IV
STUDYING
Build Your Own Data Science Journey

CHAPTER IV: ROADMAP

Self-Taught vs.
Bootcamps vs. DS
Degree
So, you've made up your mind to pursue a career
in data science. Now the question comes: how do
you get the training you need? The options might
seem overwhelming: should you go for a
traditional degree, sprint through a bootcamp, or
take the DIY route with self-teaching?

While all three paths can lead to a successful

career in data science, there are notable
differences between them. DS degree programs,
for example, provide a thorough foundation in
data science theory and principles, but may not
provide as much practical experience in
programming as bootcamps. Bootcamps offer
intense, focused training in a short period of time,
but may not provide the same depth of knowledge
as a DS degree. Self-taught programmers have the
freedom to choose their own learning path and
pace, but may struggle with accountability and
structure.

Page 50
Build Your Own Data Science Journey

CHAPTER IV: ROADMAP

There is always an argument that it is hard to find a job as a self-taught data scientist or
a bootcamp graduate, but in my experience, that is not always the case. I am self-taught,
and most of the data scientists I know are either bootcamp graduates or have a degree
in CS or physics, not DS. Of course, it is easier to get a job with a DS degree, but the
degree is not a requirement. Every time you see "Data Science, Mathematics, or
Statistics or other relevant degree required" bullet in a job description, just ignore it.
They always want a degree, for every job, in any sector. Not having a degree does not
disqualify you. The only thing you really need is the courage to apply for that job despite
deeming yourself unqualified.

So now, as we established that you could get a data scientist position (Yes, even at
FAANG) as a self-taught programmer or bootcamp graduate, not only as a DS degree
graduate, let's figure out which route would be actually better for you.

The Degree Route

Going for a degree in data science or a related
field like statistics, computer science, or applied
mathematics is the traditional path. Universities
offer structured curriculums, expert faculty, and
the prestige of a formal qualification. Having a
degree can also open doors to research roles and
is often preferred for positions in academia or
more traditional corporate settings.

However, getting a degree is a significant time

commitment and often comes with a hefty price
tag. In the U.S., you'll end up paying more than
$40k a year, even with government support and
all. It may be overkill if you're looking to switch
careers quickly or if you already have some of the
foundational knowledge.

But if you haven't earned a bachelor's degree yet,

it's generally recommended to go the traditional
route and aim for a degree in a STEM (Science,
Technology, Engineering, Mathematics) field.

Page 51
Build Your Own Data Science Journey

CHAPTER IV: ROADMAP

Data science is interdisciplinary, and a strong foundation in mathematics, statistics, and
computer science will give you the tools you need for a robust career. If you already
have a bachelor's degree, even if it's in a field unrelated to data science, you have more
flexibility. In this case, bootcamps and self-study become more viable options, but each
has its pros and cons.

If you're thinking about a master's or a PhD, remember that the reputation of the school
matters—a lot. Grad degrees are expensive and time-consuming, so if you commit to
one, aim for a reputable institution. A mediocre school may not give you the return on
investment you're hoping for. In this game, brand names can give you an edge,
especially when networking or competing for top-tier roles.

Do not get me wrong, I am in no way bashing the DS degree route; if you have the time
and resources to do it, of course, go for it! You will definitely benefit from it. But if you do
not, consider a bootcamp or a self-taught route.

The Bootcamp Route

If you do not have much time to switch careers and you are a beginner, you can pick the
bootcamp route. The DS bootcamp will give you a foundation in 4-6 months and add a
couple of projects to your portfolio. You will still have to put a lot of extra work into your
skills and portfolio to get that Junior DS position, but it will be faster than getting a
degree.

Bootcamps have gained popularity as a more fast-paced, hands-on alternative to

traditional degrees. The downside is that bootcamps can be expensive and their rapid
pace is not for everyone. Some employers may also be skeptical about the depth of your
knowledge if your education is solely based on a bootcamp, although this perception is
gradually changing.

Bootcamps offer a quicker path into the field, but they come with their own set of
considerations. First off, not all bootcamps are created equal. There are some high-
quality ones out there, but they tend to be expensive. On the flip side, there are also
many subpar bootcamps that might not give you the skills you need. Do your research
thoroughly.

Moreover, bootcamps usually offer foundational knowledge, the basics you could
potentially learn on your own. They're excellent for discipline and structure, so if you
struggle with self-motivation but don’t want to invest in a graduate degree, a good
bootcamp could be a sensible choice. In conclusion, bootcamp is a great choice if you
want to pick up fundamentals fast and have the resources to afford a good bootcamp.
The Self-Taught Route Hybrid Approaches
If you do not have much resources but Don't forget, you're not locked into any
have good discipline, you can pick the single path. Many successful data
self-taught route. There are a lot of free scientists use a mix of these approaches.
websites, courses, and books to For example, you might have a degree in
familiarize yourself with any technology a different field and then fill in your data
out there. If that's not enough and you science gaps with a bootcamp or self-
need specific direction or structure, there study. Or maybe you start with self-study
are a lot of paid courses where to test the waters and then commit to a
instructors take you on a journey from formal degree program to deepen your
beginner to whatever stage you want to knowledge.
be.
Choosing between a degree, a bootcamp,
One of the biggest challenges that self- or self-study isn't a one-size-fits-all
taught data scientists face is the lack of decision. Consider your own learning
structure and accountability that comes style, your career goals, and yes, your
with traditional education. Without a set financial situation as well. Each path has
curriculum or deadlines, it can be easy to its merits and drawbacks, but all can lead
get sidetracked or lose motivation. To to a rewarding career in data science.
overcome this, self-taught programmers Whatever route you pick, remember that
should consider setting clear goals for in the ever-evolving world of data science,
themselves, creating a study schedule, the learning never really stops.
and finding a community of peers or
mentors for support and guidance.

So, self-study requires a tremendous

amount of discipline, but it's the most
flexible and cost-effective route. If you
already have a bachelor's or master's
degree in a non-STEM field and are highly
motivated, self-study can work well.
However, it's not easy. You'll need to
create your own curriculum, seek out
resources, and set—and stick to—a study
schedule. Also, lacking formal credentials
means you'll have to prove your skills in
other ways, like building a strong portfolio
or obtaining certifications. Focus on your
portfolio, make it stand out, and you will
get your first DS job!

Page 53
Do I need a PhD to get a data science job?
Probably not. Alternative routes like a master's
degree, bootcamps, or self-studying
Yes, a fair number of data science can provide the necessary skills for
job posts require a PhD. The PhD these jobs.
requirement is usually a signal from
the company that the position is While having a PhD can offer
considered senior. But in many specific advantages, such as
cases, years of practical work can specialized knowledge and research
be just as valuable as, if not more skills, it is by no means a universal
so than, a doctoral degree. requirement for a career in data
science. If you already have a
For those just starting out in data bachelor's or master's degree and
science, a PhD is generally not are considering your next steps,
necessary and might even be know that there are multiple paths
considered excessive for entry-level to success in this field. The most
positions. In these roles, your suitable path for you will depend on
skillset and practical abilities are your career goals, the time and
often more critical than having an resources you have available, and
advanced degree. your personal learning preferences.

Page 54
How Long to Learn Data Science
from Zero to Job-Ready?
For self-learners, a frequent The next phase involves branching
question is how long it takes to out into more complex subjects.
become qualified for an internship This could include machine
or a Junior Data Scientist role. A learning, data visualization, and
typical timeframe is 6 to 12 months, potentially even specialized areas,
but there are some essential details depending on individual interests.
to consider. During this phase, the goal is to
learn various aspects of data
First, the 6 to 12-month estimate science and begin tackling more
assumes many hours of advanced projects.
concentrated study every day. So
it's like a full-time job. In the Towards the end of the 6 to 12-
beginning, the focus should be on month timeframe, focusing on
laying a solid foundation. This particular topics of interest or
involves mastering a programming market demand may be beneficial.
language commonly used in data This is also an opportune time to
science, such as Python, and getting compile a portfolio of projects
a handle on basic statistical showcasing acquired skills and
concepts. With full-time study, these knowledge.
initial objectives are generally
achievable within a few months. To sum it up, a 6 to 12-month
period of dedicated, daily focus can
be a reasonable expectation for
becoming competent in data
science through self-study.
However, if other obligations make
this level of commitment unfeasible,
the journey may require a longer
investment of time.

Page 55
 LEARNING DATA SCIENCE

WHILE WORKING

FULL TIME
So, you're holding down a full-time job and you're interested in the field of
data science. You've heard the buzzwords—machine learning, data
visualization, big data—and you're eager to join in. But there's a question
you may have: "Is it possible to learn data science while working full-time?"
The short answer is yes, but there are important considerations to keep in
mind.

Becoming proficient in data science is not something that will happen

overnight or even over a few months. We're talking about a commitment
that will likely span over many years, especially if you're starting from
ground zero.

Given the time constraints of a full-time job, weekends are invaluable. These
uninterrupted stretches can be incredibly helpful for deep focus, allowing
you to delve into complex subjects without constantly checking the time.
Reserve blocks of 3-4 hours on Saturdays and Sundays specifically for your
data science studies. During this time, turn off social media notifications,
inform your family and friends that you'll be busy, and immerse yourself in
learning.

Beyond weekends, consistency is key. Commit to spending at least 1-2

hours each weekday on your data science journey. These could be hours
early in the morning before work, during lunch breaks, or late at night when
the rest of the world is winding down.
Page 56
Learning Data Science While Working
Full-Time

The point is to make it a habit. Habits, once formed, are self-sustaining. The
daily grind ensures that you are continually reinforcing and building upon
what you've learned, not losing the thread between one study session and
the next.

Since time is a constraint, you need to be efficient in your learning.

Structure your learning path in advance to ensure you're not wandering
aimlessly through topics.

Despite best efforts, there will be days when work pressures will make it
impossible to dedicate time to learning. That's okay. Life happens, and the
journey to becoming a data scientist is a marathon, not a sprint. The
important thing is to get back on track as soon as you can.

Learning data science while working full-time is definitely challenging but

entirely achievable. With strategic planning, time management, and
dedication, you can set yourself on the path to a new career. Use the
template on the next page to set your learning goals.

Page 57
SETTING YOUR GOALS

30 DAYS
ACTION PLAN

60 DAYS
ACTION PLAN

90 DAYS
ACTION PLAN
How to Get Out of (or Completely Avoid)
the Tutorial Trap
Online tutorials and courses are great. From platforms like Coursera to Udemy, you've
got the world's best instructors right at your fingertips, often for free or a modest fee.
But there's a hidden danger lurking in this abundance: the "Tutorial Trap." This term
refers to the cycle of taking endless courses or reading countless tutorials, feeling like
you're gaining valuable knowledge but never actually applying it.

Completing online tutorials gives you knowledge, sure. But it doesn't automatically give
you the skill to apply that knowledge independently. This disconnect between knowledge
and application can create an "illusion of competence." You feel you've mastered a
subject when, in reality, you've only scratched the surface.

Let's say you complete a machine learning course on Udemy. You earn a certificate and
then get suggestions for three more courses. You complete those too. Earn more
certificates. Rinse and repeat. The issue? You haven't actually applied what you've
learned. Until you've applied knowledge to real-world problems, hit roadblocks, and
found solutions, you won't truly master your skills.

Change the Examples! If a course teaches sentiment analysis using a Twitter dataset,
don't just copy the example. Apply the same principles to a new dataset—like Amazon
reviews. This forces you to adapt your skills to new contexts, building your confidence
along the way. This way, you will learn how to use the same building blocks on different
kinds of problems or on different kinds of “environment”, and build something new, and
I guess I don’t need to explain how valuable that is to excel in any given career! This will
surely increase your confidence to tackle real world problems on your own.

Try to Improve the examples! It is normal for courses to use a running case study or
project to teach different concepts to the students. Instead of blindly following through
the case studies, you can use your existing knowledge to slightly improve the project by
adding additional functionalities to the project. This will help you stand out from the
crowd as well as help in building a great portfolio.

Do a Personal Project! After you finish a course, hold off on jumping into another one.
Instead, work on a personal project that applies what you've learned. Document your
process and end result in a blog post or video. Having a portfolio of such projects not
only demonstrates your skills but also builds your confidence.

Page 59
 ACTION PRIORITY

MATRIX
THE ACTION PRIORITY MATRIX IS A GREAT WAY TO VISUALISE WHAT
TASKS TAKE PRIORITY OVER OTHERS, AND HOW TO BEST ALLOCATE
YOUR TIME TOWARDS THEM.

HIGH IMPACT

CHANGE THE EXAMPLES MAJOR PROJECTS

LOW EFFORT

TUTORIALS IMPROVE THE EXAMPLES HIGH EFFORT

LOW IMPACT
How to Get Out of (or Completely Avoid)
the Tutorial Trap
Don't Try to Learn it All! In the era of information overload, it's easy to feel like you
have to learn everything. New frameworks, programming languages, and machine
learning algorithms are coming out all the time, and the fear of missing out is real. But
here's the thing: you don't need to learn it all.

Trying to cover every possible topic will not only burn you out but also deepen your
entanglement in the Tutorial Trap. The goal isn't to know everything; it's to understand
key principles deeply and know how to apply them in real-world scenarios. Specializing
in certain areas and becoming exceptionally good at them will be more beneficial in the
long run than having a superficial understanding of a wide array of topics.

Don't Be Afraid to Fail! Fear of failure can be paralyzing, but it's crucial to remember
that failure is often the stepping stone to success. In fact, one of the reasons I hesitated
to build projects on my own was a deep-rooted fear of messing up. But here's a reality
check: you will fail, and that's perfectly okay.

Learning to code—or mastering any skill, really—is a process. You won't become an
expert overnight. Each failure is a lesson learned, each mistake a stepping stone on your
path to proficiency.

Page 61
Productivity Tips
The journey to becoming proficient in data science is long and filled with
challenges, both expected and unexpected. One key factor that can
determine your success is your productivity. How efficiently and effectively
you manage your time and resources can make all the difference. Now, let’s
explore some top productivity tips specifically tailored for those learning
data science.

Consistency is Key
Creating a consistent learning schedule helps your mind get into a “learning
mode“, making it easier to focus and understand complex topics.
Time Blocking
Allocate specific blocks of time for different tasks. For example, reading
research papers from 9-10 AM and coding from 11 AM - 1 PM.
Pomodoro Technique
25 minutes of focused learning followed by a 5-minute break can
significantly increase your productivity and decrease burnout.
Project Management Tool
Whether it's ClickUp, Notion, or good old-fashioned Post-its, keeping track
of tasks, and deadlines is crucial.
Pareto Principle
Focus on the 20% of topics that will give you 80% of the needed skills. For
example, concentrate on mastering Python libraries like Pandas and
Matplotlib before diving into more niche tools.
Learning Environment
Having a dedicated learning space can mentally prepare you to focus on
your studies.
Physical Exercise
A little bit of exercise can go a long way in keeping your mind sharp and
focused.
Don't Skimp on Sleep
Your brain needs rest to process all the new information you're learning. A
proper night's sleep can help improve memory retention and problem-
solving skills.

Page 62
 Common
Mistakes Data
Scientists
Make and How
to Avoid Them
The road to becoming a data scientist is a challenging one, filled with many
opportunities to make mistakes. While errors are an integral part of any
learning journey, some pitfalls are more common and detrimental than others.
In this chapter, we delve into frequent missteps data scientists often make,
offering you strategies on how to avoid them.

1. Trying to Learn Multiple Tools/Languages at Once.

Newcomers to the field often make the mistake of trying to learn multiple
programming languages at once, thinking it will make them more marketable.
While it may seem like a good idea, it’s more beneficial to master one language
first. If you're just starting, focus on either Python or R, but not both
simultaneously. However, you can consider pairing your main language with
SQL since they serve different purposes.

Page 63
Common Mistakes Data Scientists Make
and How to Avoid Them
2. Using Existing Code as a "Black Box"
Another common mistake is using pre-existing code without fully
understanding how it works. While it might seem efficient to use code that
someone else has already written, doing so without understanding it can
lead to problems down the line. Building projects from scratch will help
reinforce what you've learned and make you a better programmer in the
long run.

3. Lacking Consistency in Studying

Many people study intensely for a short period and then take a long break,
thinking they've learned enough. However, the skills you need to become
proficient require frequent, regular practice. Think of it like muscle memory;
the more you code, the better and faster you will become at it. Therefore,
consistent learning and practice are key to your development.

4. Not Spending Enough Time on Data Visualization

Often, data scientists focus solely on developing complex algorithms and
models, neglecting data visualization. However, a well-crafted visualization
can sometimes be more impactful than the most complicated model. They
clarify your findings and are often more accessible to stakeholders. Make it
a rule to include effective visualizations in all your projects.

5. Unclear Problem Statements

Starting a project without a clearly defined problem or question is another
common mistake. You may be tempted to dive into the data and start
building models without knowing what you’re trying to solve or prove.
Always begin with a clear problem statement, and select your methods and
tools based on how well they can answer that question.

Page 64
6. Avoiding Discussions and Competitions
Some people avoid discussions or competitive scenarios because they lack
confidence or don't want to be challenged. However, avoiding these
opportunities for peer interaction is a mistake. Engaging in dialogue,
discussions, and even competitions exposes you to different perspectives and
problem-solving approaches, thereby enriching your own understanding and
capabilities.

7. Neglecting Communication Skills

Last but not least, poor communication skills can hold you back in your data
science career. The ability to clearly explain complex data findings in a way that
non-technical stakeholders can understand is invaluable. Therefore, never
underestimate the power of strong communication skills. Work on explaining
your findings clearly, presenting your ideas effectively, and improving your
writing skills for documentation.

In the ever-changing landscape of data science, avoiding these common

mistakes can accelerate your growth and make your journey less fraught with
unnecessary struggle. By paying attention to these areas, you’ll be better
positioned to become a more effective and well-rounded data scientist.

Page 65
Build Your Own Data Science

Journey

A resume says what you've done;

a portfolio shows how you've
done it.

Chapter V
PORTFOLIO
Build Your Own Data Science Journey

CHAPTER V: PORTFOLIO

"Science" in data science can be misleading. After all, data science is a very
practical endeavor. You'll often hear that it's 80% preprocessing and 20%
modeling, but the reality is, it's 100% about understanding, learning, and
applying. True mastery comes not just from reading about algorithms or
understanding the theoretical underpinnings, but from the nitty-gritty of
coding, wrangling data, and making those algorithms work on real-world
data. Let’s dive into the practical side of building your data science portfolio.

What to Include?

Before looking for a job, you should build a portfolio of open-source

projects on Github (and a website to accompany it, if you can). I recommend
creating these projects:

Classification Project: Choose a public dataset (just not a popular one like
the Sign Language MNIST dataset) and implement a classification model.
Test multiple models and tuning parameters, and write up your results and
their interpretation.

Data Visualization Project: Create a visually compelling dashboard that

narrates a story around a socially relevant topic. This project should
highlight your skills in data storytelling.

Web Scraping Project: Develop a project that scrapes information from

various sources, like sports news or financial data. The ultimate aim could
be to automate a website that regularly updates its content based on the
data scraped.

If you are more advanced, then you should definitely play around with the
latest machine learning algorithms, like LLMs, causal forests, reinforcement
learning, or something. You have endless opportunities here. A great
resource is Papers With Code (Google it!), which links machine learning
papers with GitHub code. That is a very efficient way of finding the most
recent trends in ML.
DATA SCIENCE PROJECT
PIPELINE
5 MOST IMPORTANT STEPS

DATA COLLECTION

DATA CLEANING

MODELING

EVALUATION

DEPLOYMENT
How Many Projects Do You Where to Publish Your
Need in Your Portfolio? Portfolio?
When starting out as a Junior data When considering where to
scientist, it's advisable to have 3 to 5 showcase your data science
quality projects in your portfolio. portfolio, you essentially have three
These projects should showcase a optimal platforms, and the best part
range of skills, from data is, you can leverage all three
preprocessing to model building, simultaneously to maximize your
ensuring that each one reach and visibility:
demonstrates a clear grasp of the GitHub: This platform is a
data science process. As you gain must-have for every data
experience, it's essential to curate scientist. Not only does it allow
and refine your portfolio. Aim to you to host and share your
maintain a selection of 3 to 5 code, but it also offers a
projects, but these should reflect a collaborative environment
higher complexity and depth of where others can contribute,
understanding. Make sure to comment, or learn from your
remove or update any beginner- projects.
level projects to ensure your Personal Website: A dedicated
portfolio accurately showcases your personal website can act as a
current expertise and skill set. digital business card and a
central hub for your
professional identity. Here, you
can present your projects with
detailed write-ups, integrate
interactive visualizations, and
even provide downloadable
resources. A well-crafted
website reflects professionalism
and a commitment, setting you
apart from the competition.

Page 69
Where to Publish Your How to Come Up with
Portfolio? Project Ideas?
Blog on Medium: Medium is Dive into Your Personal
an excellent platform for data Interests: Every data scientist
scientists to not just showcase starts somewhere, and often, the
their projects, but to also tell most engaging projects stem from
the story behind them. By personal interests. Take me as an
blogging about the challenges example: I'm passionate about
faced, the methodology climbing. So, I embarked on a
adopted, or the insights derived project centered around Everest
from a project, you can engage summiters. Surprisingly, despite the
a broader audience, establish richness and uniqueness of the
your expertise, and even receive data, it appeared to be rarely used
feedback from peers in the in the data science community. This
community. allowed my project to not only
reflect my personal passion but also
Incorporating all three options stand out in the crowd. Building
provides a multi-faceted approach around your interests lets your
to establishing your presence in the character and personality shine
data science community. While through, revealing the sort of topics
GitHub demonstrates your and challenges that captivate you.
technical skills, a personal website
offers a holistic view of your Strategically Align with Your
capabilities, and Medium provides a Career Goals: It's not just about
platform to communicate the doing what you love but also about
nuances and insights of your work. showcasing your specialization and
Together, they form a powerful trio potential value to employers.
that can help you stand out in the
pull of applicants.

Page 70
How to Come Up with Project Ideas?

By researching companies or fields Start with a Curiosity: No niche?

you're aiming for, you can sculpt a No problem. Sometimes, a simple
portfolio that speaks directly to question can pave the way for a
them. For instance, if you're looking captivating project. Maybe you're
at the health tech industry, a project intrigued by shifts in gender
analyzing wearable fitness data distribution across college majors,
might be a great fit. If that industry or perhaps you're curious about the
happens to coincide with your correlation between study habits
personal interests, it's a bonus! and academic success. Whatever
the question, let your curiosity
Stay Unique: A word of caution - guide your research. Google can be
steer clear of overused datasets like your starting point in scouting data
Titanic, MNIST, or other popular sources that can help you answer
beginner-level datasets. While your research question.
they're fantastic for honing skills,
they're not a good fit for portfolios. When All Else Fails...
Your aim should be to showcase
originality and a fresh perspective. If you find yourself at a dead-end,
You can still use Kaggle datasets, devoid of ideas, remember: it's okay
just consider datasets with fewer to lean on preset projects initially,
notebooks. And remember, Kaggle especially as a beginner. The goal is
isn't the only resource. There are to start somewhere, learn
numerous data repositories and continuously, and gradually mold
APIs. Plus, you can try to scrape the your portfolio to reflect a balance of
data yourself. passion, expertise, and ingenuity.

Below, I have provided ten ideas

you can use in your projects if you
are a beginner and do not know
where even to start.

Page 71
Project ideas
If you're at the beginning of your journey, unsure of your niche, or simply
working on your initial projects and seeking inspiration, consider exploring
one of these ideas.

1. Web Scraping and NLP: Word Frequency in the Adventures of

Sherlock Holmes.

Use web scraping and NLP to find the most frequent words in classic
literature: Arthur Conan Doyle's book, The Adventures of Sherlock Holmes.
In this project, scrape a book from an html file taken from the website
Project Gutenberg (which contains a large corpus of books) using the
Python requests package and encoding it to utf-8. Here is the URL to
scrape from. Then, extract text from this web data using an HTML parser
from the BeautifulSoup library to get the text. Next, analyze the
distribution of words using the Natural Language ToolKit (nltk), a very
common NLP task used to gain insights on textual data and lay the
groundwork for deeper analysis. Find the top 10 popular words and
visualize their distribution using the dispersion_plot. Do not forget to
create a word cloud plot using WordCloud and matplotlib.
The natural language processing tools used here are required for many
important data science tasks as a vast proportion of the world's data is
unstructured and includes a great deal of text. To complete this project, you
need to know how to import web data into Python and how to work with
natural language text.

2. Statistics: Build a Logistic Regression Model in Python from

Scratch.

Building a Logistic Regression Model from scratch involves understanding

the math behind the model, then implementing the equations using
programming logic. First generate a random linearly separable dataset
using numpy. Split the dataset into features (X) and the target variable (y).

Page 72
Project ideas
Split the data into training and testing sets for model validation. Implement
the model using Sigmoid Function, Cost Function and Gradient Descent.
After training, use the logistic regression model to make predictions on the
test set. Create a confusion matrix, calculate precision, recall, and F1 score
for a comprehensive evaluation. Compare its performance (in terms of
accuracy, speed, etc.) with logistic regression from libraries like scikit-
learn. Consider implementing additional features, such as regularization, to
improve your logistic regression's capabilities.

3. Dataviz: Visualizing Stolen Animals Data.

For this data visualization project, you'll dive into the "MPS Stolen Animals
Dashboard Data" from the London Metropolitan Police Service, available
here. Using a tool of your choice, such as Tableau, Plotly, or R Shiny, create
a dashboard highlighting:
1) a map showcasing the number of stolen animals by borough,
2) a chart displaying the count of animals by type,
3) historical trends of animal thefts, and
4) data on how many stolen pets were eventually recovered.
Aim to uncover hidden insights and patterns. For reference, view London
police's official visualization here.

4. Data Analysis: Analyzing California High School SAT Performance

vs. Poverty Level.

In this project, the objective is to investigate the relationship between SAT

performance and poverty levels among high schools in California. Begin by
downloading two key datasets: the "Unduplicated Student Poverty – Free or
Reduced-Price Meals Data 2015–16" from the California Department of
Education here and the "California SAT Test Results 1516" from Kaggle here.

Page 73
Project ideas
After loading these datasets into Python using pandas, clean and
preprocess the data to ensure consistency and accuracy. The datasets
should then be merged. Delve into the core analysis by examining the
correlations between SAT reading and writing scores, and between SAT
writing and math scores. Further, explore the relationship between the
percentage of students eligible for free or reduced-price lunch (an indicator
of poverty) and their SAT scores. Visualize these relationships using tools
like plotly or matplotlib. Conclusively, run a regression to quantitatively
understand the influence of poverty levels on SAT scores. The end product
should be a well-documented Python notebook with clear visualizations and
a summary of the key insights, trends, and patterns observed.

[Link] Learning: Image Recognition of Current U.S. Congress

Members.

This is a deep learning project centered on image recognition for current

Congress members. Download the names data for Congress members
from here, and collect images of each Congress member using Google
Images Downloader. The focus of the project is primarily on faces, so utilize
MTCNN to extract and save individual faces from each image.
Employ Tensorflow to process the images, maintaining a consistent size of
224x224. For the deep learning model, leverage transfer learning. Begin
with pre-trained models like MobileNet, ResNet 50, or ResNet 101, which
have previously been trained on ImageNet. Subsequently, adapt your model
using the 'compile' method in Tensorflow, selecting an optimizer such as
'adam' and determining the appropriate loss function for categorical
prediction. After training the model, assess its accuracy and save it for
future use.
For a more specialized approach, use a ResNet50 model trained on
FaceNet to recognize different faces. Once the model is compiled and
trained, deploy MTCNN to fetch faces from new images. Lastly, compare the
performances of the two models to determine which has better accuracy in
identifying the Congress members.

Page 74
Project ideas
6. Time Series Analysis: Trends and Seasonal Patterns in Traffic
Accidents.

Dive into a time-focused exploration of UK's traffic accidents using data

from UK Road Safety: Traffic Accidents and Vehicles that can be
downloaded here. Start with an exploratory analysis, focusing on patterns
related to time, location, road characteristics, and other factors like
weather. Examine correlations between different factors such as the day of
the week or special road conditions. Then analyze accidents across various
time frames—hourly, daily, monthly, and yearly—pinpointing evident
patterns or peaks. Investigate potential seasonal fluctuations, like winter
upticks or holiday surges. Optionally, deploy forecasting models to predict
future accident occurrences. Ultimately, synthesize the insights into
actionable recommendations for enhancing road safety in the UK.

7. Machine Learning: Predicting Student Grades Based on Social and

Academic Factors.

In this project, explore this dataset on Kaggle that deals with student
alcohol consumption and its relationship to academic and social factors
that can be found here. Your goal is to use various machine learning
models like Logistic Regression, Decision Trees, Random Forest, and
Support Vector Machines to predict the final grades based on features like
alcohol consumption, study time, family relationships, and extracurricular
activities. After building and training your models, you'll compare their
performance using metrics like accuracy, F1-score, and ROC AUC to
determine the most effective model for this specific dataset. This project will
not only improve your skills in data preprocessing, model selection, and
evaluation but also give you a nuanced understanding of how different
algorithms perform on the same dataset.

Page 75
Project ideas
8. Data Cleaning: Profitable App Profiles for the App Store and
Google Play Markets.

This project aims to identify app types that attract the most users since
revenue is driven by in-app ads in free apps. Use datasets from Google Play
Store from here and Mobile App Store data from here, focusing on manual
data cleaning without Numpy or Pandas.
Start by cleaning the data: remove incorrect entries, duplicates, non-English
apps, and non-free apps. The analysis is divided into different categories.
First, assess the most common apps by genre. Next, shift focus to the most
popular apps in both markets. The App Store analysis relies on the average
user ratings for genre determination, while Google Play offers clearer
insights with install numbers. However, pay attention to vague install
number ranges. Delve deeper into genres that exhibit high potential. Try to
identify the most promising niche for a free app.

9. Causal Machine Learning: Impact of Reemployment Bonuses on

Unemployment Duration.

In this project, tap into the Pennsylvania Reemployment Bonus

Demonstration dataset (1988-1989) curated by Mathematica Policy
Research and available here. The dataset was created to assess the
efficiency of varying bonus amounts in speeding up reemployment and
lessening the benefits availed by unemployment insurance claimants. Your
goal is to pinpoint the causal impact of these reemployment bonuses on
unemployment duration, UI benefits, and subsequent employment and
earnings.

Page 76
Project ideas
Begin with a rudimentary regression model. Examine the relationship
between log(𝑖𝑛𝑢𝑖𝑑𝑢𝑟1) and 𝑡4 without introducing any controls. This offers
an initial glance at the potential impact. Calculate both the Average
Treatment Effect (ATE) and the Average Treatment Effect on the Treated
(ATT) for group 4's reemployment bonus on unemployment duration. Next,
harness the Causal ML package to dive deeper.
Implement the DML algorithm to extract causal relationships, ensuring that
the algorithm controls for confounding factors effectively. Make use of
various algorithms present within the Causal ML package. This step not
only allows you to draw more precise inferences but also to compare the
efficiency and effectiveness of different algorithms in the causal analysis
realm. End the project with a thorough assessment of your findings,
underlining the nuances of causal inference you gleaned during the
analysis.

10. Sentiment Analysis: Depression Detection through Twitter.

Diving into the vast world of Twitter, this project aims to discern signs of
depression using Natural Language Processing (NLP). Gather tweets using
Twitter API. Evenly split between those indicating depression and those that
don't. Tweets flagged for depression are obtained by querying depression-
linked keywords like "depression", "antidepressants", and "anxiety". Next, do
a manual review to filter out tweets using these terms in non-depressive
contexts.
Do model training—include Logistic Regression, SVM, and Deep Neural
Networks with varying embeddings. Use the Vader tool to differentiate
sentiment distribution in tweets and dive deep into emoji usage patterns,
drawing a clearer picture of users' emotional states. Coupled with visual
word clouds and emoji bar charts, the analysis offers a comprehensive lens
into the linguistic indicators of depression on Twitter.

Page 77
How to manage your portfolio
projects
Managing your projects can be What sets them apart from the rest
tough. The biggest problem many is that they decided to show their
face is finishing a project and then projects even if they were unsure.
being brave enough to show it to They also had doubts, worried
others. Some doubt themselves, about what others would say, and
thinking their project isn't good felt their work wasn't the best. But
enough. Others might think, "It's they went ahead, made a portfolio,
just one project, I should have more and got jobs.
before publishing," or worry it looks
too simple and people will criticize If you're having a hard time finishing
it. Others try to build projects based a project, ask for help. If your ideas
on too complex ideas, that seem are too complex, simplify them, pick
more fitting for a PhD dissertation what you can do, and work on that.
than a simple portfolio piece. If you're worried about sharing a
Consequently, they never see it project that's not perfect,
through to the end. Also, when hit a remember that only you know its
coding problem, some give up too full story. Maybe add a part
soon. explaining the problems you faced
and what you learned. This shows
Sound familiar? The truth is that you can learn and adjust. And yes,
perfectionism can be the enemy of it's okay to start with just one
progress. The best advice? Just go project. Once it's out there, you'll
ahead and publish your project likely find the motivation to add
anyway. Hopefully, recognising that more.
these anxieties are universal brings
some comfort. Remember, many
successful data scientists once had
these same insecurities.

Page 78
How to manage your portfolio
projects
Showcase your work across various Here's the truth: As soon as you
platforms, from GitHub and began learning to code, you got
personal portfolio websites to what you needed to start a coding
Medium articles and LinkedIn. Take project. Learning by doing is one of
that step and put yourself out the most effective approaches. If
there. you have too many ideas, just pick
one and go for it.
Another popular problem many
face is the initial hesitation to Don't let fear hold you back. Start
embark on a project. Thoughts like today!
"Am I smart enough?", "Am I skilled
enough?", or "Should I complete a
few more courses before starting?"
often plague the mind. Some even
believe their project idea is boring,
while others create an
overwhelming list of potential
projects, unsure of where to begin.

Page 79
 PORTFOLIO

Do & Dont's
DO

Create a website to
showcase your portfolio

Use dataviz to make

results more digestible.

DON'T

Don't include too many

projects

Don't overcomplicate
your project ideas

@polinacsv
Build Your Own Data Science

Journey

VI

The future depends on what you

do today.

Chapter VI
FIND A JOB
 JUNIOR DATA SCIENTIST JOB READINESS

CHECKLIST
IF YOU MEET 8 OUT OF THE 10 CRITERIA ON THE CHECKLIST, YOU ARE
CONSIDERED PREPARED TO APPLY.

CRITERIA

01 Are you proficient in at least one programming language, such as Python or R?

02 Are you familiar with core machine learning algorithms, evaluation metrics, and tools?

03 Have you completed and documented at least 2-3 data science projects?

04 Can you comfortably write basic SQL queries for data extraction?

05 Are you able to explain technical concepts to non-technical stakeholders?

06 Are you prepared for coding tests or technical assessments?

07 Have you uploaded projects to GitHub with clear READMEs and well-commented code?

08 Can you explain the difference between supervised and unsupervised learning?

09 Do you know how to diagnose and address overfitting in a machine learning model?

10 Are you familiar with version control systems, specifically Git?

REFLECTION NOTES
 HOW TO

FIND A JOB

Before you start looking for a job, it's good to think about what you're good
at and what you really want. Do you love machine learning, or are you really
good at statistical analysis? For example, if you've done a lot of work on NLP
projects, you might want to look for specific jobs at places like OpenAI or
Grammarly. Also, think about where you'd like to work: a fast-moving small
company like DataRobot or a big company like projects, you might want to
look for specific jobs at places like OpenAI or Grammarly. Also, think about
where you'd like to work: a fast-moving small company like DataRobot or a
big company like Google?Google?

While many people look for jobs on popular sites like LinkedIn and Indeed,
don't forget about specialized job sites like DataCamp Jobs Board, DataJobs,
and [Link]. Also, check the job pages of companies you're interested in
—places like NVIDIA or Palantir sometimes have unadvertised job listings.
Working with job agencies that focus on tech jobs, like Harnham, might also
help you find unique job offers.

Social media sites like Twitter and Reddit are also places where people talk
about job openings. You can watch for posts about data science jobs to find
opportunities. Plus, getting to know people—at online events, conferences,
or through school contacts—can help a lot. Sometimes, these connections
can tell you about jobs that aren't advertised.

Finding a job in data science can be tough. If you've sent out a lot of
applications but only had a few interviews, don't get discouraged.
Sometimes things like computerized resume checks or too many applicants
can make it hard. But remember, every challenge is a chance to learn and
Page 83
get better. Take any feedback you get, make changes, and keep trying.
Resume Formatting Tips
Length and Format: Limit your resume to one page. Save as a PDF if
digital, or use resume paper if printing.
Be concise: Opt for bullet points or brief sentences.
Font and Margins: Choose a font size between 10 to 12 points. Ensure
margins are at least 0.5 inches on all sides.
Pronoun Usage: Avoid first-person pronouns, especially "I".
Tense: Use the past tense.
Punctuation: Maintain consistency. If using periods at the end of bullet
points, apply them uniformly.
Formatting: Utilize bold and italics strategically to enhance readability.
Use bold and black-grey contrast to highlight things you want to catch
the eye first (like projects in the example below).
Colors: Stick to black and white. Avoid colors.
Template: Use a straightforward template. Avoid fancy designs.
Cleanliness: Ensure a neat layout. Stick to one font if possible.
Links: Always include your GitHub. If applying in the US, add LinkedIn
link. If you have them, incorporate the portfolio website and Medium
links in the header preceding your GitHub. Ensure all links are clickable.
Summary: Omit a summary section unless the resume lacks
substantial content.
Visual Elements: Avoid visual effects or decorations.
Photos: If applying within the US, do not include a photo.
Order: Adopt a reverse chronological order. Reorder sections based on
their significance. If you've worked for a renowned company, list the
company's name first, followed by your title. Otherwise, reverse the
order.
Language Proficiency: Only mention languages directly pertinent to
the position. Avoid mentioning too many programming languages.
Skills: Refrain from listing soft skills. Prioritize technical skills; they're
essential. Exclude older degrees and high school diplomas. Don't forget
to list your data science certifications.
Specialized Content: If possible, tailor your resume to a niche. For
instance, if experienced in creating recommendation engines,
emphasize that skill like it is done in the example below. From the
example, you can quickly grasp that the person is skilled at building
recommendation engines, has experience working with social media
data, and has domain expertise in business.
Resume Example

Page 85
Cover Letters
While a resume provides an overview of your qualifications and
experiences, a cover letter paints a more holistic picture of who you are. It
goes beyond the bullet points, diving into your passion, personality, and fit
for the specific role. A resume lays out the factual roadmap of your
professional journey. In contrast, the cover letter delves deeper, offering
hiring managers a glimpse into your character, motivations, and aspirations.
It's an opportunity to showcase your unique value proposition and how it
aligns with the company's mission and needs. If your resume doesn’t follow
a linear trajectory or if there are gaps, a cover letter becomes the bridge
that connects the dots. It allows you to weave a narrative that demonstrates
how varied experiences have collectively equipped you for the role in
question.

Demonstrate your genuine interest in the company by mentioning your

familiarity with their products, services, or core values. Having taken the
time to visit their "About" webpage or having firsthand experience with their
offerings can set you apart from the crowd. Though not always mandatory,
submitting a cover letter when given the option can be a game-changer.
Some hiring managers might even perceive the absence of a cover letter as
a lack of genuine interest.

A cover letter should ideally range between three-quarters to a full page.

While it's tempting to address every aspect of the job description, zero in on
your strongest attributes. If a role requires four distinct skills and you're
particularly adept at two, emphasize those. Avoid the pitfall of apologizing
for or over-explaining the skills you might lack. A well-crafted cover letter
can help you stand out in a competitive job market. It's your personal pitch,
spotlighting your most valuable assets in relation to the role. So, while it
might be an "optional" document in the application process, its potential
impact on your job prospects is profound.

Page 86
Cold Messaging
Even with a meticulously crafted The earlier, the better. As
resume and an impressive portfolio, soon as you've submitted an
the journey of job applications can online application, reach out.
sometimes lead to a silent void. The immediacy shows genuine
You've likely clicked the 'submit' interest and enthusiasm. Also,
button on numerous job portals, you will forget to reach out if will
only to be met with silence. This not do it immediately.
'black hole' effect of online job
applications, where your effort Tailor Your Message. Begin by
seemingly vanishes into oblivion, mentioning your recent
can be discouraging. But there's an application and express
underrated, proactive approach to genuine excitement about the
break this cycle: cold messaging on role. Highlighting the alignment
professional platforms like LinkedIn. between your past experiences
and the company’s goals or
Like many of you, I've experienced values demonstrates a deeper
the frustration of companies not understanding and keenness.
responding after online job Don't forget to include a link to
applications. However, during my your portfolio.
last job search, a significant shift
occurred. Nearly 75% of the initial Quantity and Quality. While
interviews I secured were a direct it's essential to reach out to
result of cold messaging on multiple professionals, ensure
LinkedIn. To craft a message that every message is personalized.
will be effective, you need to Remember, while many might
remember a few things: not respond, a single positive
reply can pave the way to an
Identify the Right Person. At interview.
the company where you're Indeed, not every message will get a
applying, try to find a data response. And that's okay. It's a
scientist, an alum from your numbers game, with persistence
university, or a recruiter. If you and consistency being key. The
can, find a data scientist in the beauty of cold messaging lies in its
department/team the job ability to bridge the gap between
position is for. If you are bold you and potential employers
enough, you can even message directly, bypassing the automated
or email the team lead. systems and crowded inboxes.
Cold Messaging
Template 1:
Subject: Exploring Junior Data Scientist Opportunities
Hi [Recipient's Name],
I hope this message finds you well. I am reaching out to express my interest
in the Junior Data Scientist opportunities at [Company Name]. With a solid
foundation in Python, R, and SQL, and a passion for turning data into
actionable insights, I believe I can contribute effectively to your team.
I recently completed a project on [briefly describe a relevant project or
accomplishment], which further solidified my skills in data cleaning,
visualization, and statistical modeling. I am eager to continue learning and
growing within a dynamic team like yours.
I have attached my resume and a link to my portfolio for your consideration:
[Link to Portfolio]. Please let me know if there's an opportunity for us to
discuss how my background and skills might align with your team’s needs.
Thank you for considering my application. I look forward to the possibility of
contributing to [Company Name].
Warm regards, [Your Name]

Template 2:
Subject: Aspiring Data Scientist Interested in Internship Opportunities
Hello [Recipient's Name],
I came across your profile while researching data-driven companies and
was impressed by [Company Name]'s innovative approach to [specific field
or project]. As a recent graduate with a degree in [Your Degree], I am keen
to apply my academic knowledge in a practical, real-world setting.
During my studies, I developed a strong analytical skill set, with hands-on
experience in machine learning, data analysis, and creating informative
visualizations. I have compiled a portfolio showcasing some of my projects,
which you can view here: [Link to Portfolio].
I am interested in learning about any internship opportunities within your
team and would be thrilled to bring my enthusiasm for data science to
[Company Name]. May I ask for the best way to apply or someone I could
speak to for further information?
Thank you for your time and assistance.
Best regards, [Your Name]

Page 88
Prepare for Interviews
Master the Interview Essentials
Here's a roadmap to prepare for different types of data science interviews:
Technical Interviews: Focus on the basics—statistics, machine
learning algorithms, and coding. You can expect questions on data
manipulation, data analysis, and problem-solving.
Behavioral Interviews: Prepare stories that showcase your
experience, teamwork, and problem-solving abilities. The STAR
(Situation, Task, Action, Result) method is an excellent framework for
structuring these stories.
Case Studies: Some interviews may require you to solve real-world
business problems on the spot. Practicing case studies beforehand can
be immensely beneficial.
Take-Home Assignments: These are designed to evaluate your
practical skills. Always adhere to the given guidelines and aim to go the
extra mile to make your work stand out.

Identify Your Knowledge Gaps

While preparing, you'll inevitably encounter areas where your

understanding is less than solid. Do not overlook these knowledge gaps;
instead, use them as a roadmap for focused learning. Here's how:
Self-Assessment: Periodically test yourself on key areas of data
science, from programming languages like Python or R to machine
learning algorithms and statistical concepts. Use this exercise to identify
your weak points.
Peer Reviews: Having a friend or colleague conduct a mock interview
can provide an outsider’s perspective on your areas for improvement.
Online Resources: Utilize platforms that offer specialized quizzes or
assessments focused on data science topics. Websites like
StrataScratch, Interview Query, or LeetCode can be helpful in this
regard.

Page 89
Prepare for Interviews
Identify Your Knowledge Gaps
Craft a Study Plan
Once you've identified your weak zones, the next step is to tackle them
head-on. Create a study plan that allocates specific time to each subject
matter, combining theoretical reading with practical exercises. This focused
approach helps you turn weak zones into strong points, further bolstering
your readiness for interviews.
Tech Up
If you find gaps in your technical skills, consider enrolling in specialized
courses or workshops. Platforms like DataCamp, DataQuest, Coursera offer
courses tailored for various skill levels and specific data science topics.
These can be particularly effective in addressing targeted areas where you
lack expertise.
Mock Interviews
Simulate real interview settings by participating in mock interviews, either
with friends in the field or through professional services. This practice not
only helps you get comfortable with the interview format but also allows you
to receive constructive feedback in real-time.

By meticulously preparing for interviews and actively seeking to close your

knowledge gaps, you're not just improving your chances of landing the job
but also becoming a better data scientist. Every interview is an opportunity
to learn, grow, and get one step closer to becoming the professional you
aspire to be.

How to Pass an Interview

Interviews can be tough. You have to convince people you don't know that
you're right for a job you've only read a short description about. In a short
time, you might be asked many questions about things you may or may not
know. You also have to figure out if you'd like to work at that company. But
with some planning, interviews can be easier and less scary.

Page 90
Prepare for Interviews
How to Pass an Interview
The first interview is usually talking to a recruiter. It's important to make a
good first impression here. This person will decide if you should talk to
more people at the company. They'll ask you about your past jobs, what you
know, and why you want this job. Be honest and show you're interested in
the job.

It's also a good time to ask questions. You can ask about the job, the
company, and what they expect. Sometimes they might talk about your
salary expectations. It's best to wait and talk about this later. At the end of
this interview, ask what is the next step in the interview process.

After your initial chat with the recruiter, if things go well, you'll move on to
more detailed interviews. These interviews can be different types and will
help the company get to know you better and see if you're a good fit for the
job.

Here are some of the interview types you might have:

Technical Interview: This is where your data science acumen will be

put to the test. Expect to be presented with scenarios, datasets, or
algorithmic challenges. You might be asked to design experiments,
discuss machine learning models, or dissect data problems. Brush up
on core concepts and be prepared to explain your thought process.
Behavioral Interview: This interview will probe how you handle real-
world challenges, teamwork, and setbacks. Given the collaborative
nature of data science, where you often work with cross-functional
teams, it's essential to demonstrate strong communication skills and a
positive approach to problem-solving.

Page 91
Prepare for Interviews
How to Pass an Interview
Meet-the-Team Interview: Data science doesn't operate in a vacuum.
This interaction lets you meet potential colleagues, whether they're
fellow data scientists, analysts, or members from other departments like
engineering or marketing. It’s an opportunity to gauge team dynamics
and showcase how you’d contribute to collaborative projects.
Interview with Upper Management: Engaging with senior leaders or
managers will offer insight into the company's strategic direction and
the role of data science within it. This is your moment to align with the
company's vision, showcase your potential contributions, and
understand future growth opportunities.

Besides these interviews, there's usually a test or case study you need to
complete. It is usually distributed in one of these forms :
Online Tech Assessment: An online challenge, often timed, testing your
data manipulation, modeling, and coding skills. You might face real-
world data problems, statistical tests, or coding exercises.
Take-Home Tech Assessment: Instead of a timed online challenge, you
might be given a more extensive project to complete at home. This
could involve data cleaning, exploratory data analysis, and/or building a
predictive model. This format allows for deeper analysis and is a chance
to showcase your data science skills.

For many, the Online Tech Assessment or the Take-Home Tech Assessment
can be the most nerve-wracking part of the interview process. These
assessments are designed to test your practical knowledge, problem-
solving abilities, and technical skills in a more direct manner than
conversational interviews. But with early preparation, and the right
resources, you can approach these with confidence.

Page 92
Prepare for Interviews
How to Pass an Interview
Several platforms can help you practice and refine your technical skills:
Strata Scratch: This is especially useful for data science interviews. It
offers datasets and problems that mirror real-world challenges in data
science.
Interview Query: Another great resource for data science questions. It
provides problems sourced from actual tech interviews.
HackerRank: From my experience, companies often use this platform to
administer tech assessments. It's beneficial to practice here because it
allows you to get used to the interface, ensuring there are no surprises
on the actual test day.
LeetCode: While this is more geared towards software engineering
roles, it's a fantastic place to sharpen your coding and algorithmic skills.

Many individuals, perhaps due to overconfidence or misjudgment, believe

that merely possessing experience will naturally make them shine in an
interview. This is a misconception I've both felt and observed in others.
When faced with probing questions like, "Tell us about a time you had
trouble communicating," a lack of preparation can cause even the most
experienced individuals to ramble or go off track.

Preparation not only aids in delivering concise and relevant answers but
ensures that your vast experience is presented in the most impactful
manner. Without the right kind of preparation, you might miss the chance
to truly showcase your capabilities.

Remember, interviews are two-way streets. While the company evaluates

you, it's equally vital for you to assess the company's fit for your aspirations
and career goals. Good luck on your data science job hunt!

Page 93
 INTERVIEW PREPAREDNESS

CHECKLIST
TASKS

01 Research the company: Understand its history, products, culture, and recent news.

Understand the job role: Familiarize yourself with the responsibilities and required skills
02 from the job description.

Prepare your answers: Anticipate and practice responses for common interview
03 questions.

Know the logistics: Confirm interview details (date, time, venue) and test any required
04 technology for virtual interviews.

05 Prepare questions: Think of meaningful questions to ask your interviewer.

Practice soft skills: Work on clear communication, eye contact, and positive body
06 language.

07 Mindset and relaxation: Engage in relaxation techniques and get a good night's sleep.

08 Plan your follow-up: Be ready to send a thank-you note after the interview.

APPROACH YOUR INTERVIEW WITH PREPARATION AND AUTHENTICITY.

GOOD LUCK!
Salary Negotiations
Before you plunge into the new A lukewarm response might raise
role, it’s essential to pause, evaluate doubts about your commitment. If,
the offer, and potentially negotiate. however, the offer significantly
Your excitement should not cloud deviates from your expectations, it's
your judgment, and immediately okay to voice your concerns
accepting an offer without tactfully. Referring to industry
contemplation may not serve your standards or your own experience
best interests. can serve as a good benchmark.
But be careful not to use this tactic
It’s critical to meticulously review for minor adjustments; it’s more
every detail of the offer. Even if the suitable when the disparity is vast
position seems perfect, there’s enough to be a deal-breaker.
often room for negotiation.
Remember, the interim between A valid written offer should detail
receiving an offer and accepting it is your title, salary, benefits package,
when you wield the most leverage. stock options, if any, and any other
Employers have invested time in the details. It should also provide a
recruitment process, and now that reasonable timeframe for your
they've chosen you, they’re eager to decision. If this window is too short,
onboard you. This eagerness arises like less than a week or even 24
not only from their belief in your fit hours, consider it a red flag. Such
but also from the costs associated pressure tactics aim to induce quick
with prolonged recruitment. decisions, often ridden with anxiety
and impulsiveness. If you're
Typically, the job offer, either awaiting outcomes from other
conveyed via a call or email, comes interviews, it’s courteous and
from the recruiter, hiring manager, beneficial to inform these
or HR personnel. Regardless of who employers about your existing offer.
extends the offer, express your This might expedite their decision-
gratitude and enthusiasm. making if they’re keen on hiring you.

Page 95
Salary Negotiations
Now, onto the negotiation itself. When it comes to negotiating a job
Many shy away from this step, offer, the first thing that often
perceiving it as confrontational or springs to mind is the salary. Before
fearing it might come off as even reaching the final stages of
ungrateful, especially if the offer your interviews, it's a good idea to
already exceeds their current look up the typical salaries for data
compensation. However, scientists. Don't just stick to a
negotiations are an integral part of general search—delve into specifics
the hiring process. Your salary related to the industry, the city, and
should reflect industry standards, the company you're considering.
and not just your previous earnings While the offer might seem
or the company's initial proposal. generous compared to your current
With the diverse roles within data or previous role, it's essential to
science, such as data analyst, ensure that it aligns with what your
machine learning engineer, or data peers are earning at the same
engineer, salary benchmarks are company.
still evolving, leading to significant
discrepancies. It's not uncommon While you're going through the
for two equally qualified individuals interview process, remember that
to earn vastly different salaries you don't have to disclose your
based purely on their job title. current salary if asked. If you're
probed about your current earnings
As you venture into negotiations, or your salary expectations, it might
arm yourself with industry insights, be best to avoid giving a direct
understand your worth, and answer. By revealing too much, you
prioritize your needs. Your risk the company making a lower
commitment to maximizing your offer, thinking they've already met
offer today will pave the way for your expectations.
better prospects in your future
endeavors. Remember, you've
earned this chance; now, it's about
making the most of it.

Page 96
Salary Negotiations
It's crucial to understand that you Your capacity to negotiate can vary.
shouldn't feel guilty or hesitant to Having another job offer in hand
ask for what you believe is a fair can significantly bolster your
salary. Don't base this on what position. It tells the company there's
you've earned in the past. Instead, demand for your skills at a higher
focus on what you want in this new price elsewhere. Similarly, if you're
role. Many companies initially content in your current role or
present a lower offer, expecting enjoy better perks, you're less likely
candidates to negotiate. It's to jump ship for a lower offer.
commonplace for candidates to
secure at least a 5% bump from the Lastly, ensure you view the offer in
initial offer. its entirety. Add up all the additional
benefits; things like health
Yet, there's more to an offer than insurance and 401k matches can
just the salary. Consider other add substantial value. Always look
direct financial benefits: signing at the bigger picture, not just the
bonuses, relocation allowances, or salary.
stock options. For instance, a
company might find it easier to And remember, it's extremely rare
offer a one-time signing bonus than for companies to retract offers
to adjust the base salary. simply because you've chosen to
negotiate. If they do, it's probably
Some benefits, like healthcare plans not a place you'd want to work.
or 401k matches, might be Negotiating is a standard part of the
standardized across a company, hiring process, and doing so
making them harder to negotiate. respectfully is entirely within your
But other perks, such as flexible right. If a company penalizes you for
working hours, educational that, consider it a bullet dodged.
allowances, or conference budgets,
might be up for discussion.

Page 97
 Imposter
Syndrome

Taking a major step in your career journey, especially in a challenging field like
data science, can sometimes lead to feelings of imposter syndrome. It's that
uneasy feeling where you question your own achievements and secretly fear
being unveiled as a "pretender".

Data science, with its vastness, often amplifies this sentiment. If you ask
around, a data scientist might be seen as an analyst to some, a statistician to
others, or even an engineer, visualizer, or business expert. These are all
specialized roles in their own right. With the rapid expansion and evolution of
the field, there's an underlying pressure to keep up with the latest
advancements.

Given these challenges, it's not surprising if sometimes it feels overwhelming.

But here's a perspective shift that's helped me: instead of comparing myself to
a "perfect data scientist", I focus on the richness of my own unique journey. I've
come to terms with the fact that it's impossible to grasp everything in data
science – there's always a new algorithm, tool, or language coming up. And
guess what? That's okay. No one knows it all.

Don't let self-doubt overshadow your worth. If it ever feels like you're alone in
this feeling, know that many of us grapple with these emotions, given the
expansive nature of the field.

Page 98
Recognizing and Avoiding Burnout in
Data Science
Burnout can be a silent intruder in any profession, and data science is no
exception. This chapter aims to help you recognize the early signs of
burnout and offers strategies to keep it at bay.

1. Understanding Burnout

Burnout isn’t just feeling tired on a Monday morning; it's a deep-seated

sense of weariness that persists over time. It's when the challenges of the
job no longer seem invigorating but rather overwhelming. A blend of
physical exhaustion, emotional cynicism, and a sense of reduced personal
accomplishment characterizes it. In the context of data science, it might
manifest as a reluctance to tackle new data problems, a feeling of
detachment from projects, or a belief that your contributions don't matter.

2. Early Warning Signs

Spotting burnout early is crucial for effective intervention. Signs might

include:
Decreased Enthusiasm: Projects and challenges that once excited you
now seem burdensome.
Increased Irritability: You find yourself having shorter patience with
colleagues or getting agitated over minor issues.
Physical Fatigue: Feeling consistently drained or falling ill frequently.
Reduced Performance: Tasks that once took an hour now drag on for
several, or you struggle with problems you previously handled with
ease.

Page 99
Recognizing and Avoiding Burnout in
Data Science
3. Strategies to Avoid Burnout

Set Clear Boundaries: Know when to switch off. Ensure you have a
defined work schedule and stick to it. Avoid late nights and working on
weekends.
Take Regular Breaks: Every hour, take a 5-10 minute break. Stretch,
take a walk, or do something non-work related.
Continuous Learning in Moderation: While staying updated is
essential, don’t pressurize yourself to learn everything immediately.
Allocate specific times in your week dedicated to learning.
Seek Support: Talk to colleagues, join data science communities, or
find a mentor. Sharing your feelings or challenges can provide relief.
Practice Self-care: Engage in activities you love outside of work.
Prioritize Tasks: Not everything on your to-do list is urgent. Learn to
differentiate between what’s essential and what can wait.

4. What If You're Already Burned Out?

If you're already feeling the intense weight of burnout, it's essential to act.
Seek Professional Help: Sometimes, talking to a counselor or
therapist can help navigate feelings of extreme burnout.
Consider a Vacation: Taking time away from work can provide much-
needed relief and perspective.
Re-evaluate Your Role: Perhaps it's time to consider if your current
role aligns with your personal and professional goals. Maybe a change
in position, department, or even organization is in order.
In conclusion, while the realm of data science is undoubtedly challenging,
it's crucial to prioritize mental and emotional well-being. Recognizing
burnout and taking proactive steps to avoid it can ensure a long, fulfilling
career in the field. Remember, it's not about how long you work but how
effectively you can work.

Page 100
WRAPPING UP!
Alright, I promise, these are the final words you'll read in this guide. First
and foremost, thank you for making it to the end! I hope the insights shared
in this guide will be beneficial for your career path.

I have the last but still very important advice for you: Some people feel
disappointed if they don't get the title of 'data scientist' as their first job in
the field. But, believe me, that's not an indicator of failure! Before getting a
'data scientist' job, I embraced various roles for over two years - research
assistant, research analyst, data consultant, and data analyst. And guess
what? Most of these roles involved building machine learning models, which
is at the heart of data science.

The essence of the data scientist role is embedded in many titles. Whether
you're labeled as a data analyst, research analyst, or something else, it's the
work and dedication that counts. Every role is a stepping stone, especially if
you're keen to learn and expand your horizons. So don't get too hung up on
titles. Explore, learn, and grow!

Lastly, and truly the final piece of news: If you ever find yourself with
questions or need guidance tailored to your journey, don't hesitate to reach
out to me on Instagram. And for those who've been waiting, yes, I've taken
the leap to offer mentorship support. If you're interested, get in touch to
learn about the mentorship services available.

Once again, thank you for your time and trust. Here's to the exciting,
fulfilling, and ever-evolving journey of data science! Stay passionate and
remember – you've got this!

Page 101
THANK YOU
Contact me :
@polinacsv
support@[Link]

[Link]