Build a Career in Data Science

Build a Career in Data Science

Emily Robinson Jacqueline Nolis

Copyright

For online information and ordering of this and other Manning books, please visit www.manning.com. The publisher offers discounts on this book when ordered in quantity.

For more information, please contact

          Special Sales Department

          Manning Publications Co.

          20 Baldwin Road

          PO Box 761

          Shelter Island, NY 11964

          Email: orders@manning.com

©2020 by Emily Robinson and Jacqueline Nolis. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by means electronic, mechanical, photocopying, or otherwise, without prior written permission of the publisher.

Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks. Where those designations appear in the book, and Manning Publications was aware of a trademark claim, the designations have been printed in initial caps or all caps.

Recognizing the importance of preserving what has been written, it is Manning’s policy to have the books we publish printed on acid-free paper, and we exert our best efforts to that end. Recognizing also our responsibility to conserve the resources of our planet, Manning books are printed on paper that is at least 15 percent recycled and processed without the use of elemental chlorine.

Development editor: Karen Miller

Review editor: Ivan Martinović

Production editor: Lori Weidert

Copy editor: Kathy Simpson

Proofreader: Melody Dolab

Typesetter: Dennis Dalinnik

Cover designer: Leslie Haimes

ISBN: 9781617296246

Printed in the United States of America

Dedication

From Emily, to Michael, and From Jacqueline, to Heather, Amber, and Laura, for the love and support you provided us throughout this journey.

Brief Table of Contents

Copyright

Brief Table of Contents

Table of Contents

Preface

Acknowledgments

About This Book

About the Authors

About the Cover Illustration

1. Getting started with data science

Chapter 1. What is data science?

Chapter 2. Data science companies

Chapter 3. Getting the skills

Chapter 4. Building a portfolio

2. Finding your data science job

Chapter 5. The search: Identifying the right job for you

Chapter 6. The application: Résumés and cover letters

Chapter 7. The interview: What to expect and how to handle it

Chapter 8. The offer: Knowing what to accept

3. Settling into data science

Chapter 9. The first months on the job

Chapter 10. Making an effective analysis

Chapter 11. Deploying a model into production

Chapter 12. Working with stakeholders

4. Growing in your data science role

Chapter 13. When your data science project fails

Chapter 14. Joining the data science community

Chapter 15. Leaving your job gracefully

Chapter 16. Moving up the ladder

 Epilogue

 Appendix. Interview questions

Index

List of Figures

List of Tables

Table of Contents

Copyright

Brief Table of Contents

Table of Contents

Preface

Acknowledgments

About This Book

About the Authors

About the Cover Illustration

1. Getting started with data science

Chapter 1. What is data science?

1.1. What is data science?

1.1.1. Mathematics/statistics

1.1.2. Databases/programming

1.1.3. Business understanding

1.2. Different types of data science jobs

1.2.1. Analytics

1.2.2. Machine learning

1.2.3. Decision science

1.2.4. Related jobs

1.3. Choosing your path

1.4. Interview with Robert Chang, data scientist at Airbnb

What was your first data science journey?

What should people look for in a data science job?

What skills do you need to be a data scientist?

Summary

Chapter 2. Data science companies

2.1. MTC: Massive Tech Company

2.1.1. Your team: One of many in MTC

2.1.2. The tech: Advanced, but siloed across the company

2.1.3. The pros and cons of MTC

2.2. HandbagLOVE: The established retailer

2.2.1. Your team: A small group struggling to grow

2.2.2. Your tech: A legacy stack that’s starting to change

2.2.3. The pros and cons of HandbagLOVE

2.3. Seg-Metra: The early-stage startup

2.3.1. Your team (what team?)

2.3.2. The tech: Cutting-edge technology that’s taped together

2.3.3. Pros and cons of Seg-Metra

2.4. Videory: The late-stage, successful tech startup

2.4.1. The team: Specialized but with room to move around

2.4.2. The tech: Trying to avoid getting bogged down by legacy code

2.4.3. The pros and cons of Videory

2.5. Global Aerospace Dynamics: The giant government contractor

2.5.1. The team: A data scientist in a sea of engineers

2.5.2. The tech: Old, hardened, and on security lockdown

2.5.3. The pros and cons of GAD

2.6. Putting it all together

2.7. Interview with Randy Au, quantitative user experience researcher at Google

Are there big differences between large and small companies?

Are there differences based on the industry of the company?

What’s your final piece of advice for beginning data scientists?

Summary

Chapter 3. Getting the skills

3.1. Earning a data science degree

3.1.1. Choosing the school

3.1.2. Getting into an academic program

3.1.3. Summarizing academic degrees

3.2. Going through a bootcamp

3.2.1. What you learn

3.2.2. Cost

3.2.3. Choosing a program

3.2.4. Summarizing data science bootcamps

3.3. Getting data science work within your company

3.3.1. Summarizing learning on the job

3.4. Teaching yourself

3.4.1. Summarizing self-teaching

3.5. Making the choice

3.6. Interview with Julia Silge, data scientist and software engineer at RStudio

Before becoming a data scientist, you worked in academia; how have the skills learned there helped you as a data scientist?

When deciding to become a data scientist, what did you use to pick up new skills?

Did you know going into data science what kind of work you wanted to be doing?

What would you recommend to people looking to get the skills to be a data scientist?

Summary

Chapter 4. Building a portfolio

4.1. Creating a project

4.1.1. Finding the data and asking a question

4.1.2. Choosing a direction

4.1.3. Filling out a GitHub README

4.2. Starting a blog

4.2.1. Potential topics

4.2.2. Logistics

4.3. Working on example projects

4.3.1. Data science freelancers

4.3.2. Training a neural network on offensive license plates

4.4. Interview with David Robinson, data scientist

How did you start blogging?

Are there any specific opportunities you have gotten from public work?

Are there people you think would especially benefit from doing public work?

How has your view on the value of public work changed over time?

How do you come up with ideas for your data analysis posts?

What’s your final piece of advice for aspiring and junior data scientists?

Summary

Chapters 1–4 resources

Books

Blog posts

2. Finding your data science job

Chapter 5. The search: Identifying the right job for you

5.1. Finding jobs

5.1.1. Decoding descriptions

5.1.2. Watching for red flags

5.1.3. Setting your expectations

5.1.4. Attending meetups

5.1.5. Using social media

5.2. Deciding which jobs to apply for

5.3. Interview with Jesse Mostipak, developer advocate at Kaggle

What recommendations do you have for starting a job search?

How can you build your network?

What do you do if you don’t feel confident applying to data science jobs?

What would you say to someone who thinks I don’t meet the full list of any job’s required qualifications?

What’s your final piece of advice to aspiring data scientists?

Summary

Chapter 6. The application: Résumés and cover letters

6.1. Résumé: The basics

6.1.1. Structure

6.1.2. Deeper into the experience section: generating content

6.2. Cover letters: The basics

6.2.1. Structure

6.3. Tailoring

6.4. Referrals

6.5. Interview with Kristen Kehrer, data science instructor and course creator

How many times would you estimate you’ve edited your résumé?

What are common mistakes you see people make?

Do you tailor your résumé to the position you’re applying to?

What strategies do you recommend for describing jobs on a résumé?

What’s your final piece of advice for aspiring data scientists?

Summary

Chapter 7. The interview: What to expect and how to handle it

7.1. What do companies want?

7.1.1. The interview process

7.2. Step 1: The initial phone screen interview

7.3. Step 2: The on-site interview

7.3.1. The technical interview

7.3.2. The behavioral interview

7.4. Step 3: The case study

7.5. Step 4: The final interview

7.6. The offer

7.7. Interview with Ryan Williams, senior decision scientist at Starbucks

What are the things you need to do to knock an interview out of the park?

How do you handle the times where you don’t know the answer?

What should you do if you get a negative response to your answer?

What has running interviews taught you about being an interviewee?

Summary

Chapter 8. The offer: Knowing what to accept

8.1. The process

8.2. Receiving the offer

8.3. Negotiation

8.3.1. What is negotiable?

8.3.2. How much you can negotiate

8.4. Negotiation tactics

8.5. How to choose between two good job offers

8.6. Interview with Brooke Watson Madubuonwu, senior data scientist at the ACLU

What should you consider besides salary when you’re considering an offer?

What are some ways you prepare to negotiate?

What do you do if you have one offer but are still waiting on another one?

What’s your final piece of advice for aspiring and junior data scientists?

Summary

Chapter 5–8 resources

Books

Blog posts and courses

3. Settling into data science

Chapter 9. The first months on the job

9.1. The first month

9.1.1. Onboarding at a large organization: A well-oiled machine

9.1.2. Onboarding at a small company: What onboarding?

9.1.3. Understanding and setting expectations

9.1.4. Knowing your data

9.2. Becoming productive

9.2.1. Asking questions

9.2.2. Building relationships

9.3. If you’re the first data scientist

9.4. When the job isn’t what was promised

9.4.1. The work is terrible

9.4.2. The work environment is toxic

9.4.3. Deciding to leave

9.5. Interview with Jarvis Miller, data scientist at Spotify

What were some things that surprised you in your first data science job?

What are some issues you faced?

Can you tell us about one of your first projects?

What would be your biggest piece of advice for the first few months?

Summary

Chapter 10. Making an effective analysis

10.1. The request

10.2. The analysis plan

10.3. Doing the analysis

10.3.1. Importing and cleaning data

10.3.2. Data exploration and modeling

10.3.3. Important points for exploring and modeling

10.4. Wrapping it up

10.4.1. Final presentation

10.4.2. Mothballing your work

10.5. Interview with Hilary Parker, data scientist at Stitch Fix

How does thinking about other people help your analysis?

How do you structure your analyses?

What kind of polish do you do in the final version?

How do you handle people asking for adjustments to an analysis?

Summary

Chapter 11. Deploying a model into production

11.1. What is deploying to production, anyway?

11.2. Making the production system

11.2.1. Collecting data

11.2.2. Building the model

11.2.3. Serving models with APIs

11.2.4. Building an API

11.2.5. Documentation

11.2.6. Testing

11.2.7. Deploying an API

11.2.8. Load testing

11.3. Keeping the system running

11.3.1. Monitoring the system

11.3.2. Retraining the model

11.3.3. Making changes

11.4. Wrapping up

11.5. Interview with Heather Nolis, machine learning engineer at T-Mobile

What does machine learning engineer mean on your team?

What was it like to deploy your first piece of code?

If you have things go wrong in production, what happens?

What’s your final piece of advice for data scientists working with engineers?

Summary

Chapter 12. Working with stakeholders

12.1. Types of stakeholders

12.1.1. Business stakeholders

12.1.2. Engineering stakeholders

12.1.3. Corporate leadership

12.1.4. Your manager

12.2. Working with stakeholders

12.2.1. Understanding the stakeholder’s goals

12.2.2. Communicating constantly

12.2.3. Being consistent

12.3. Prioritizing work

12.3.1. Both innovative and impactful work

12.3.2. Not innovative but still impactful work

12.3.3. Innovative but not impactful work

12.3.4. Neither innovative nor impactful work

12.4. Concluding remarks

12.5. Interview with Sade Snowden-Akintunde, data scientist at Etsy

Why is managing stakeholders important?

How did you learn to manage stakeholders?

Was there a time where you had difficulty with a stakeholder?

What do junior data scientists frequently get wrong?

Do you always try to explain the technical part of the data science?

What’s your final piece of advice for junior or aspiring data scientists?

Summary

Chapters 9–12 resources

Books

Blogs

4. Growing in your data science role

Chapter 13. When your data science project fails

13.1. Why data science projects fail

13.1.1. The data isn’t what you wanted

13.1.2. The data doesn’t have a signal

13.1.3. The customer didn’t end up wanting it

13.2. Managing risk

13.3. What you can do when your projects fail

13.3.1. What to do with the project

13.3.2. Handling negative emotions

13.4. Interview with Michelle Keim, head of data science and machine learning at Pluralsight

When was a time you experienced a failure in your career?

Are there red flags you can see before a project starts?

How does the way a failure is handled differ between companies?

How can you tell if a project you’re on is failing?

How can you get over a fear of failing?

Summary

Chapter 14. Joining the data science community

14.1. Growing your portfolio

14.1.1. More blog posts

14.1.2. More projects

14.2. Attending conferences

14.2.1. Dealing with social anxiety

14.3. Giving talks

14.3.1. Getting an opportunity

14.3.2. Preparing

14.4. Contributing to open source

14.4.1. Contributing to other people’s work

14.4.2. Making your own package or library

14.5. Recognizing and avoiding burnout

14.6. Interview with Renee Teate, director of data science at HelioCampus

What are the main benefits of being on social media?

What would you say to people who say they don’t have the time to engage with the community?

Is there value in producing only a small amount of content?

Were you worried the first time you published a blog post or gave a talk?

Summary

Chapter 15. Leaving your job gracefully

15.1. Deciding to leave

15.1.1. Take stock of your learning progress

15.1.2. Check your alignment with your manager

15.2. How the job search differs after your first job

15.2.1. Deciding what you want

15.2.2. Interviewing

15.3. Finding a new job while employed

15.4. Giving notice

15.4.1. Considering a counteroffer

15.4.2. Telling your team

15.4.3. Making the transition easier

15.5. Interview with Amanda Casari, engineering manager at Google

How do you know it’s time to start looking for a new job?

Have you ever started a job search and decided to stay instead?

Do you see people staying in the same job for too long?

Can you change jobs too quickly?

What’s your final piece of advice for aspiring and new data scientists?

Summary

Chapter 16. Moving up the ladder

16.1. The management track

16.1.1. Benefits of being a manager

16.1.2. Drawbacks of being a manager

16.1.3. How to become a manager

16.2. Principal data scientist track

16.2.1. Benefits of being a principal data scientist

16.2.2. Drawbacks of being a principal data scientist

16.2.3. How to become a principal data scientist

16.3. Switching to independent consulting

16.3.1. Benefits of independent consulting

16.3.2. Drawbacks of independent consulting

16.3.3. How to become an independent consultant

16.4. Choosing your path

16.5. Interview with Angela Bassa, head of data science, data engineering, and machine learning at iRobot

What’s the day-to-day life as a manager like?

What are the signs you should move on from being an independent contributor?

Do you have to eventually transition out of being an independent contributor?

What advice do you have for someone who wants to be a technical lead but isn’t quite ready for it?

What’s your final piece of advice to aspiring and junior data scientist?

Summary

Chapters 13–16 resources

Books

Blogs

 Epilogue

 Appendix. Interview questions

A.1. Coding and software development

A.1.1. FizzBuzz

A.1.2. Tell whether a number is prime

A.1.3. Working with Git

A.1.4. Technology decisions

A.1.5. Frequently used package/library

A.1.6. R Markdown or Jupyter Notebooks

A.1.7. When should you write functions or packages/libraries?

A.1.8. Example manipulating data in R/Python

A.2. SQL and databases

A.2.1. Types of joins

A.2.2. Loading data into SQL

A.2.3. Example SQL query

A.2.4. Example SQL query continued

A.2.5. Data types

A.3. Statistics and machine learning

A.3.1. Statistics terms

A.3.2. Explain p-value

A.3.3. Explain a confusion matrix

A.3.4. Interpreting regression models

A.3.5. What is boosting?

A.3.6. Favorite algorithm

A.3.7. Training vs. test data

A.3.8. Feature selection

A.3.9. Deploying a new model

A.3.10. Model behavior

A.3.11. Experimental design

A.3.12. Flaws in experimental design

A.3.13. Bias in sampled data

A.4. Behavioral

A.4.1. Project that had the most impact

A.4.2. Data surprises

A.4.3. Previous job reflections

A.4.4. Senior person making a mistake based on data

A.4.5. Disagreements with teammates

A.4.6. Difficult problems

A.5. Brain teasers

A.5.1. Estimation

A.5.2. Combinatorics

Index

List of Figures

List of Tables

Preface

How do I get your job?

As veteran data scientists, we’re constantly being asked this question. Sometimes, we’re asked directly; at other times, people ask indirectly through questions about the decisions we’ve made in our careers to get where we are. Under the surface, the people asking the questions seem to have a constant struggle, because so few resources are available for finding out how to become or grow as a data scientist. Lots of data scientists are looking for help with their careers and often not finding clear answers.

Although we’ve written blog posts with tactical advice on how to handle specific moments in a data science job, we’ve struggled with the lack of a definitive text covering the end-to-end of starting and growing a data science career. This book was written to help these people—the thousands of people who hear about data science and machine learning but don’t know where to start, as well as those who are already in the field and want to understand how to move up.

We were happy to get this chance to collaborate in creating this book. We both felt that our respective backgrounds and viewpoints complemented each other and created a better book for you. We are

Jacqueline Nolis— I received a BS and MS in mathematics and a PhD in operations research. When I started working, the term data science didn’t yet exist, and I had to figure out my career path at the same time that the field was defining itself. Now I’m a consultant, helping companies grow data science teams.

Emily Robinson— I got my undergraduate degree in decision sciences and my master’s in management. After attending a three-month data science bootcamp in 2016, I started working in data science, specializing in A/B testing. Now I work as a senior data scientist at Warby Parker, tackling some of the company’s biggest projects.

Throughout our careers, we’ve both built project portfolios and experienced the stress of adjusting to a new job. We’ve felt the sting of being rejected for jobs we wanted and the triumph of seeing our analyses positively affect the business. We’ve faced issues with a difficult business partner and benefited from a supportive mentor. Although these experiences taught us so much in our careers, to us the true value comes from sharing them with others.

This book is meant to be a guide to career questions in data science, following the path that a person will take in the career. We start with the beginning of the journey: how to get basic data science skills and understand what jobs are actually like. Then we go through getting a job and how to get settled in. We cover how to grow in the role and eventually how to transition up to management—or out to a new company. Our intention is for this book to be a resource that data scientists continue to go back to as they hit new milestones in their careers.

Because the focus on career is very important for this book, we chose to not focus deeply on the technical components of data science; we don’t cover topics such as how to choose the hyperparameters of a model or the minute details of Python packages. In fact, this book doesn’t include a single equation or line of code. We know that plenty of great books out there cover these topics; we wanted instead to discuss the often-overlooked but equally important nontechnical knowledge needed to succeed in data science.

We included many personal experiences from respected data scientists in this book. At the end of each chapter, you’ll find an interview describing how a real, human data scientist personally handled dealing with the concepts that the chapter covers. We’re extremely happy with the amazing, detailed, and vulnerable responses we got from all the data scientists we talked to. We feel that the examples they provide from their lives can teach much more than any broad statement we might write.

Another decision we made in writing this book was to make it opinionated. By that, we mean we intentionally chose to focus on the lessons we’ve learned as professional data scientists and by talking to others in the community. At times, we make statements not everyone might agree with, such as suggesting that you should always write a cover letter when applying for jobs. We felt that the benefit of providing viewpoints that we strongly believe are helpful to data scientists was more important than trying to write something that contained only objective truths.

We hope that you find this book to be a helpful guide as you progress in your data science career. We’ve written it to be the document we wish we had when we were aspiring and junior data scientists; we hope that you’ll be glad to have it now.

Acknowledgments

First and foremost, we’d like to thank our spouses, Michael Berkowitz and Heather Nolis. Without them, this book would not have been possible (and not just because Michael wrote the first draft of some of the sections despite being a bridge professional and not a data scientist, or because Heather evangelized half of the machine learning engineering content).

Next, we want to acknowledge the staff at Manning who guided us through this process, improved the book, and made it possible in the first place. Thank you especially to our editor, Karen Miller, who kept us on track and coordinated all the various moving parts.

Thank you to all the reviewers who read the manuscript at various points and provided invaluable detailed feedback: Brynjar Smári Bjarnason, Christian Thoudahl, Daniel Berecz, Domenico Nappo, Geoff Barto, Gustavo Gomes, Hagai Luger, James Ritter, Jeff Neumann, Jonathan Twaddell, Krzysztof Jędrzejewski, Malgorzata Rodacka, Mario Giesel, Narayana Lalitanand Surampudi, Ping Zhao, Riccardo Marotti, Richard Tobias, Sebastian Palma Mardones, Steve Sussman, Tony M. Dubitsky, and Yul Williams. Thank you as well to our friends and family members who read the book and offered their own suggestions: Elin Farnell, Amanda Liston, Christian Roy, Jonathan Goodman, and Eric Robinson. Your contributions helped shape this book and made it as helpful to our readers as possible.

Finally, we want to thank all of our end-of-chapter interviewees: Robert Chang, Randy Au, Julia Silge, David Robinson, Jesse Mostipak, Kristen Kehrer, Ryan Williams, Brooke Watson Madubuonwu, Jarvis Miller, Hilary Parker, Heather Nolis, Sade Snowden-Akintunde, Michelle Keim, Renee Teate, Amanda Casari, and Angela Bassa. Additionally, we’re grateful for those who contributed to sidebars throughout the book and suggested interview questions for the appendix: Vicki Boykis, Rodrigo Fuentealba Cartes, Gustavo Coelho, Emily Bartha, Trey Causey, Elin Farnell, Jeff Allen, Elizabeth Hunter, Sam Barrows, Reshama Shaikh, Gabriela de Queiroz, Rob Stamm, Alex Hayes, Ludamila Janda, Ayanthi G., Allan Butler, Heather Nolis, Jeroen Janssens, Emily Spahn, Tereza Iofciu, Bertil Hatt, Ryan Williams, Peter Baldridge, and Hlynur Hallgrímsson. All these people provided valuable perspectives, and together, they know much more than we ever could.

About This Book

Build a Career in Data Science was written to help you enter the field of data science and grow your career in it. It walks you through the role of a data scientist, how to get the skills you need, and the steps to getting a data science job. After you have a job, this book helps you understand how to mature in the role and eventually become a larger part of the data science community, as well as a senior data scientist. After reading this book, you should be confident about how to advance your career.

Who should read this book

This book is for people who have not yet entered the field of data science but are considering it, as well as people who are in the first few years of the role. Aspiring data scientists will learn the skills they need to become data scientists, and junior data scientists will learn how to become more senior. Many of the topics in the book, such as interviewing and negotiating an offer, are worthwhile resources to come back to throughout any data science career.

How this book is organized: a roadmap

This book is broken into four parts, arranged in the chronological order of a data science career. Part 1 of the book, Getting started with data science, covers what data science is and what skills it requires:

Chapter 1 introduces the role of a data scientist and the different types of jobs that share that title.

Chapter 2 presents five example companies that have data scientists and shows how the culture and type of each company affects the data science positions.

Chapter 3 lays out the different paths a person can take to get the skills needed to be a data scientist.

Chapter 4 describes how to create and share projects to build a data science portfolio.

Part 2 of the book, Finding your data science job, explains the entire job search process for data science positions:

Chapter 5 walks through the search for open positions and how to find the ones worth investing in.

Chapter 6 explains how to create a cover letter and résumé and then adjust them for each job you apply for.

Chapter 7 provides details on the interview process and what to expect from it.

Chapter 8 is about what to do after you receive an offer, focusing on how to negotiate it.

Part 3 of the book, Settling into data science, covers the basics of the early months of a data science job:

Chapter 9 lays out what to expect in the first few months of a data science job and shows you how to make the most of them.

Chapter 10 walks through the process of making analyses, which are core components of most data science roles.

Chapter 11 focuses on putting machine learning models into production, which is necessary in more engineering-based positions.

Chapter 12 explains how to communicate with stakeholders—a task that data scientists have to do more than most other technical roles.

Part 4 of the book, Growing in your data science role, covers topics for more seasoned data scientists who are looking to continue to advance their careers:

Chapter 13 describes how to handle failed data science projects.

Chapter 14 shows you how to become part of the larger data science community through activities such as speaking and contributing to open source.

Chapter 15 is a guide to the difficult task of leaving a data science position.

Chapter 16 ends the book with the roles data scientists can get as they move up the corporate ladder.

Finally, we have an appendix of more than 30 interview questions, example answers, and notes on what the question is trying to assess and what makes a good answer.

People who haven’t been data scientists before should start at the beginning of the book, whereas people who already are in the field may begin with a later chapter to guide them in a challenge they’re currently facing. Although the chapters are ordered to flow like a data science career, they can be read out of order according to readers’ needs.

The chapters end with interviews of data scientists in various industries who discuss how the topic of the chapter has shown up in their career. The interviewees were selected due to their contributions to the field of data science and the interesting journeys they followed as they became data scientists.

liveBook discussion forum

Purchase of Build a Career in Data Science includes free access to a private web forum run by Manning Publications where you can make comments about the book, ask technical questions, and receive help from the author and from other users. To access the forum, go to https://livebook.manning.com/#!/book/build-a-career-in-data-science/discussion. You can also learn more about Manning's forums and the rules of conduct at https://livebook.manning.com/#!/discussion.

Manning’s commitment to our readers is to provide a venue where a meaningful dialogue between individual readers and between readers and the author can take place. It is not a commitment to any specific amount of participation on the part of the author, whose contribution to the forum remains voluntary (and unpaid). We suggest you try asking the author some challenging questions lest their interest stray! The forum and the archives of previous discussions will be accessible from the publisher’s website as long as the book is in print.

About the Authors

Emily Robinson

WRITTEN BY JACQUELINE NOLIS

Emily Robinson is a brilliant senior data scientist at Warby Parker and previously worked at DataCamp and Etsy.

I first met Emily at Data Day Texas 2018, when she was one of the few people who attended my talk on data science in industry. At the end of my speech, she shot her hand up and asked a great question. To my surprise, an hour later we had swapped; I was watching her calmly and casually give a great presentation while I was eagerly waiting to raise my hand and ask her a question. That day, I knew she was a hard-working and clever data scientist. A few months later, when it came time for me to find someone to co-author a book, she was at the top of my list. When I sent her the email asking whether she would be interested, I figured that there was a good chance she would say no; she was probably out of my league.

Working with Emily on this book has been a joy. She is deeply thoughtful about the struggles of junior data scientists and has the ability to clearly understand what is important. She is constantly getting her work done and somehow also is able to squeeze out extra blog posts while doing it. Now having seen her at more conferences and social events, I’ve watched as she’s talked to many data scientists and made all of them feel comfortable and welcome. She’s also an expert in A/B testing and experimentation, but it’s clear that this just happens to be the area she’s working in at the moment; she could pick up any other part of data science and be an expert in that if she wanted to.

My only disappointment is that I’m writing these words about her at the end of creating the book, and with us finishing, someone besides me will have the next opportunity to collaborate with her.

Jacqueline Nolis

WRITTEN BY EMILY ROBINSON

Whenever someone asks me whether I would recommend writing a book, I always say, Only if you do it with a co-author. But that’s not actually the full picture. It should be Only if you do it with a co-author who is as fun, warm, generous, smart, experienced, and caring as Jacqueline. I’m not sure what it’s like working with a normal co-author, because Jacqueline has always been amazing, and I feel incredibly lucky to have gotten to work with her on this project.

It would be easy for someone as accomplished as Jacqueline to be intimidating. She has a PhD in industrial engineering, got $100,000 for winning the third season of the reality television show King of the Nerds, was a director of analytics, and started her own successful consulting firm. She’s spoken at conferences across the country and is regularly asked back by her alma mater to advise math undergraduates (her major) on careers. When she spoke at an online conference, the compliments about her presentation flooded the chat, such as the best so far, excellent presentation, really helpful, and great, dynamic presentation. But Jacqueline never makes anyone feel inferior or bad for not knowing something; rather, she loves making difficult concepts accessible, such as in her great presentation called Deep learning isn’t hard, I promise.

Her personal life is equally impressive: she has a wonderfully vibrant house in Seattle with her wife, son, two dogs, and three cats. I’m hoping that she might also one day adopt a certain co-author to fill out the very few empty spaces. She and her wife, Heather, have even given a presentation to a packed audience of 1,000 people eager to hear about how they used R to deploy machine learning models to production at T-Mobile. They also possibly have the best meet-cute story of all time: they met on the aforementioned show King of the Nerds, where Heather was also a competitor.

I’m very thankful to Jacqueline, who could have earned much more money for much less aggravation by doing anything other than writing this book with me. It is my hope that our work encourages aspiring and junior data scientists to become contributors to our community who are as great as Jacqueline is.

About the Cover Illustration

Saint-Sauver

The figure on the cover of Build a Career in Data Science is captioned Femme de l'Aragon, or Aragon Woman. The illustration is taken from a collection of dress costumes from various countries by Jacques Grasset de Saint-Sauveur (1757–1810), titled Costumes de Différents Pays, published in France in 1797. Each illustration is finely drawn and colored by hand. The rich variety of Grasset de Saint-Sauveur’s collection reminds us vividly of how culturally apart the world’s towns and regions were just 200 years ago. Isolated from each other, people spoke different dialects and languages. In the streets or in the countryside, it was easy to identify where they lived and what their trade or station in life was just by their dress.

The way we dress has changed since then and the diversity by region, so rich at the time, has faded away. It is now hard to tell apart the inhabitants of different continents, let alone different towns, regions, or countries. Perhaps we have traded cultural diversity for a more varied personal life—certainly for a more varied and fast-paced technological life.

At a time when it is hard to tell one computer book from another, Manning celebrates the inventiveness and initiative of the computer business with book covers based on the rich diversity of regional life of two centuries ago, brought back to life by Grasset de Saint-Sauveur’s pictures.

Part 1. Getting started with data science

If you do a Google search for how to become a data scientist, you’ll likely be confronted with a laundry list of skills, from statistical modeling to programming in Python through communicating effectively and making presentations. One job description might describe a role that’s close to a statistician’s, whereas another employer is looking for someone who has a master’s degree in computer science. When you look for ways to gain those skills, you’ll find options ranging from going back to school for a master’s degree to doing a bootcamp to starting to do data analysis in your current job. Put together, all these combinations of paths can feel insurmountable, especially to people who aren’t yet certain that they even want to be data scientists.

The good news is that there isn’t a single data scientist who has all these skills. Data scientists share a foundation of knowledge, but they each have their own specialties, to the point that many couldn’t swap jobs. The first part of this book is designed to help you understand what all these types of data scientists are and how to make the best decisions to start your career. By the end of this part, you should be prepared with the skills and understanding to start your job search.

Chapter 1 covers the basics of data science, including the skills you need for the job and the different types of data scientists. Chapter 2 goes into detail about the role of a data scientist at five types of companies to help you better understand what the job will be like. Chapter 3 covers the paths to getting the skills required for being a data scientist and the advantages and disadvantages of each. Finally, Chapter 4 covers how to create a portfolio of data science projects to get hands-on experience doing data science and create a portfolio to show to potential employers.

1 What is data science?

This chapter covers

·  The three main areas of data science

·  The different types of data science jobs

The sexiest job of the 21st century. The best job in America. Data scientist, a title that didn’t even exist before 2008, is now the position employers can’t hire enough of and job seekers strive to become. There’s good reason for the hype: data science is a hugely growing field, with a median base salary of more than $100,000 in the United States in 2019 (http://mng.bz/XpMp). At a good company, data scientists enjoy a lot of autonomy and are constantly learning new things. They use their skills to solve significant problems, such as working with doctors to analyze drug trials, helping a sports team pick its new draftees, or redesigning the pricing model for a widget business. Finally, as we discuss in chapter 3, there’s no one way to become a data scientist. People come from all backgrounds, so you’re not limited based on what you chose to study as an undergraduate.

But not all data science jobs are perfect. Both companies and job seekers can have unrealistic expectations. Companies new to data science may think that one person can solve all their problems with data, for example. When a data scientist is finally hired, they can be faced with a never-ending to-do list of requests. They might be tasked with immediately implementing a machine learning system when no work has been done to prepare or clean the data. There may be no one to mentor or guide them, or even empathize with the problems they face. We’ll discuss these issues in more depth in chapters 5 and 7, where we’ll help you avoid joining companies that are likely to be a bad fit for a new data scientist, and in chapter 9, where we’ll advise you on what to do if you end up in a negative situation.

On the other side, job seekers may think that there will never be a dull moment in their new career. They may expect that stakeholders will follow their recommendations routinely, that data engineers can fix any data quality issues immediately, and that they’ll get the fastest computing resources available to implement their models. In reality, data scientists spend a lot of time cleaning and preparing data, as well as managing the expectations and priorities of other teams. Projects won’t always work out. Senior management may make unrealistic promises to clients about what your data science models can deliver. A person’s main job may be to work with an archaic data system that’s impossible to automate and requires hours of mind-numbing work each week just to clean up the data. Data scientists may notice lots of statistical or technical mistakes in legacy analyses that have real consequences, but no one is interested, and they’re so overloaded with work that they have no time to try to fix them. Or a data scientist may be asked to prepare reports that support what senior management has already decided, so they may worry about being fired if they give an independent answer.

This book is here to guide you through the process of becoming a data scientist and developing your career. We want to ensure that you, the reader, get all the great parts of being a data scientist and avoid most of the pitfalls. Maybe you’re working in an adjacent field, such as marketing analytics, and wondering how to make the switch. Or maybe you’re already a data scientist, but you’re looking for a new job and don’t think you approached your first job search well. Or you want to further your career by speaking at conferences, contributing to open source, or becoming an independent consultant. Whatever your level, we’re confident that you’ll find this book helpful.

In the first four chapters, we cover the main opportunities for gaining data science skills and building a portfolio to get around the paradox of needing experience to get experience. Part 2 shows how to write a cover letter and resume that will get you an interview and how to build your network to get a referral. We cover negotiation strategies that research has shown will get you the best offer possible.

When you’re in a data science job, you’ll be writing analyses, working with stakeholders, and maybe even putting a model into production. Part 3 helps you understand what all those processes look like and how to set yourself up for success. In part 4, you’ll find strategies for picking yourself back up when a project inevitably fails. And when you’re ready, we’re here to guide you through the decision of where to take your career: advancing to management, continuing to be an individual contributor, or even striking out as an independent consultant.

Before you begin that journey, though, you need to be clear on what data scientists are and what work they do. Data science is a broad field that covers many types of work, and the better you understand the differences between those areas, the better you can grow in them.

1.1. What is data science?

Data science is the practice of using data to try to understand and solve real-world problems. This concept isn’t exactly new; people have been analyzing sales figures and trends since the invention of the zero. In the past decade, however, we have gained access to exponentially more data than existed before. The advent of computers has assisted in the generation of all that data, but computing is also our only way to process the mounds of information. With computer code, a data scientist can transform or aggregate data, run statistical analyses, or train machine learning models. The output of this code may be a report or dashboard for human consumption, or it could be a machine learning model that will be deployed to run continuously.

If a retail company is having trouble deciding where to put a new store, for example, it may call in a data scientist to do an analysis. The data scientist could look at the historical data of locations where online orders are shipped to understand where customer demand is. They may also combine that customer location data with demographic and income information for those localities from census records. With these datasets, they could find the optimal place for the new store and create a Microsoft PowerPoint presentation to present their recommendation to the company’s vice president of retail operations.

In another situation, that same retail company may want to increase online order sizes by recommending items to customers while they shop. A data scientist could load the historical web order data and create a machine learning model that, given a set of items currently in the cart, predicts the best item to recommend to the shopper. After creating that model, the data scientist would work with the company’s engineering team so that every time a customer is shopping, the new machine learning model serves up the recommended items.

When many people start looking into data science, one challenge they face is being overwhelmed by the amount of things they need to learn, such as coding (but which language?), statistics (but which methods are most important in practice, and which are largely academic?), machine learning (but how is machine learning different from statistics or AI?), and the domain knowledge of whatever industry they want to work in (but what if you don't know where you want to work?). In addition, they need to learn business skills such as effectively communicating results to audiences ranging from other data scientists to the CEO. This anxiety can be exacerbated by job postings that ask for a PhD, multiple years of data science experience, and expertise in a laundry list of statistical and programming methods. How can you possibly learn all these skills? Which ones should you start with? What are the basics?

If you’ve looked into the different areas of data science, you may be familiar with Drew Conway’s popular data science Venn diagram. In Conway’s opinion (at the time of the diagram’s creation), data science fell into the intersection of math and statistical knowledge, expertise in a domain, and hacking skills (that is, coding). This image