Becoming a Data Head: How to Think, Speak, and Understand Data Science, Statistics, and Machine Learning

ISBN-13: 9781119741749 / Angielski / Miękka / 2021 / 272 str.

Alex J. Gutman; Jordan Goldmeier

Recenzje

"Big Data, Data Science, Machine Learning, Artificial Intelligence, Neural Networks, Deep Learning...It can be buzzword bingo, but make no mistake, everything is becoming "datafied" and an understanding of data problems and the data science toolset is becoming a requirement for every business person. Alex and Jordan have put together a must read whether you are just starting your journey or already in the thick of it. They made this complex space simple by breaking down the 'data process' into understandable patterns and using everyday examples and events over our history to make the concepts relatable."- Milen Mahadevan, President of 84.51°"What I love about this book is its remarkable breadth of topics covered, while maintaining a healthy depth in the content presented for each topic. I believe in the pedagogical concept of 'Talking the Walk,' which means being able to explain the hard stuff in terms that broad audiences can grasp. Too many data science books are either too specialized in taking you down the deep paths of mathematics and coding ('Walking the Walk') or too shallow in over-hyping the content with a plethora of shallow buzzwords ('Talking the Talk'). You can take a great walk down the pathways of the data field in Alex and Jordan's without fear of falling off the path. The journey and destination are well worth the trip, and the talk."- Kirk Borne, Data Scientist and Top Worldwide Influencer in Data Science"The most clear, concise, and practical characterization of working in corporate analytics that I've seen. If you want to be a killer analyst and ask the right questions, this is for you."- Kristen Kehrer, Data Moves Me, LLC and LinkedIn Top Voices in Data Science & Analytics"THE book that business and technology leaders need to read to fully understand the potential, power, AND limitations of data science."- Jennifer L. L. Morgan, PhD, Analytical Chemist at Procter and Gamble"You've heard it before: 'We need to be doing more machine learning. Why aren't we doing more sophisticated data science work?' Data science isn't the magic unicorn that will solve all of your company's problems. Becoming a Data Head brings this idea to life by highlighting when data science is (and isn't) the right approach and the common pitfalls to watch out for, explaining it all in a way that a data novice can understand. This book will be my new 'pocket reference' when communicating complicated concepts to non-technically trained leaders."- Sandy Steiger, Director, Center for Analytics and Data Science at Miami University"Individuals and organizations want to be data driven. They say they are data driven. Becoming a Data Head shows them how to actually become data driven, without the assumption of a statistics or data background. This book is for anyone, or any organization, asking how to bring a data mindset to the whole company, not just those trained in the space."- Eric Weber, Head of Experimentation & Metrics Research, Yelp"What is keeping data science from reaching its true potential? It is not slow algorithms, lack of data, lack of computing power, or even lack of data scientists. Becoming a Data Head tackles the biggest impediment to data science success, the communication gap between the data scientist and the executive. Gutman and Goldmeier provide creative explanations of data science techniques and how they are used with clear everyday relatable examples. Managers and executives, and anyone wanting to better understand data science will learn a lot from this book. Likewise, data scientists who find it challenging to explain what they are doing will also find great value in Becoming a Data Head."- Jeffrey D. Camm, PhD, Center for Analytics Impact, Wake Forest University"Becoming a Data Head raises the level of education and knowledge in an industry desperate for clarity in thinking. A must read for those working with and within the growing field of data science and analytics."- Dr. Stephen Chambal, VP for Corporate Growth at Perduco (DoD Analytics Company)"Gutman and Goldmeier filter through much of the noise to break down complex data and statistical concepts we hear today into basic examples and analogies that stick. Becoming a Data Head has enabled me to translate my team's data needs into more tangible business requirements that make sense for our organization. A great read if you want to communicate your data more effectively to drive your business and data science team forward!"- Justin Maurer, Engineering and Data Science Manager at Google"As an aerospace engineer with nearly 15 years experience, Becoming a Data Head made me aware of not only what I personally want to learn about data science, but also what I need to know professionally to operate in a data-rich environment. This book further discusses how to filter through often overused terms like artificial intelligence. This is a book for every mid-level program manager learning how to navigate the inevitable future of data science."- Josh Keener, Aerospace Engineer and Program Manager"A must read for an in-depth understanding of data science for senior executives."- Cade Saie, Chief Data Officer"Gutman and Goldmeier offer practical advice for asking the right questions, challenging assumptions, and avoiding common pitfalls. They strike a nice balance between thoroughly explaining concepts of data science while not getting lost in the weeds. This book is a useful addition to the toolbox of any analyst, data scientist, manager, executive, or anyone else who wants to become more comfortable with data science."- Jeff Bialac, Senior Supply Chain Analyst at Kroger"Gutman and Goldmeier have written a book that is as useful for applied statisticians and data scientists as it is for business leaders and technical professionals. In demystifying these complex statistical topics, they have also created a common language that bridges the longstanding communication divide that has -- until now -- separated data work from business value."- Kathleen Maley, Chief Analytics Officer at datazuum

Spis treści

Acknowledgments xiiiForeword xxiiiIntroduction xxviiPart One Thinking Like a Data HeadChapter 1 What Is the Problem? 3Questions a Data Head Should Ask 4Why Is This Problem Important? 4Who Does This Problem Affect? 6What If We Don't Have the Right Data? 6When Is the Project Over? 7What If We Don't Like the Results? 7Understanding Why Data Projects Fail 8Customer Perception 8Discussion 10Working on Problems That Matter 11Chapter Summary 11Chapter 2 What Is Data? 13Data vs. Information 13An Example Dataset 14Data Types 15How Data Is Collected and Structured 16Observational vs. Experimental Data 16Structured vs. Unstructured Data 17Basic Summary Statistics 18Chapter Summary 19Chapter 3 Prepare to Think Statistically 21Ask Questions 22There Is Variation in All Things 23Scenario: Customer Perception (The Sequel) 24Case Study: Kidney-Cancer Rates 26Probabilities and Statistics 28Probability vs. Intuition 29Discovery with Statistics 31Chapter Summary 33Part Two Speaking Like a Data HeadChapter 4 Argue with the Data 37What Would You Do? 38Missing Data Disaster 39Tell Me the Data Origin Story 43Who Collected the Data? 44How Was the Data Collected? 44Is the Data Representative? 45Is There Sampling Bias? 46What Did You Do with Outliers? 46What Data Am I Not Seeing? 47How Did You Deal with Missing Values? 47Can the Data Measure What You Want It to Measure? 48Argue with Data of All Sizes 48Chapter Summary 49Chapter 5 Explore the Data 51Exploratory Data Analysis and You 52Embracing the Exploratory Mindset 52Questions to Guide You 53The Setup 53Can the Data Answer the Question? 54Set Expectations and Use Common Sense 54Do the Values Make Intuitive Sense? 54Watch Out: Outliers and Missing Values 58Did You Discover Any Relationships? 59Understanding Correlation 59Watch Out: Misinterpreting Correlation 60Watch Out: Correlation Does Not Imply Causation 62Did You Find New Opportunities in the Data? 63Chapter Summary 63Chapter 6 Examine the Probabilities 65Take a Guess 66The Rules of the Game 66Notation 67Conditional Probability and Independent Events 69The Probability of Multiple Events 69Two Things That Happen Together 69One Thing or the Other 70Probability Thought Exercise 72Next Steps 73Be Careful Assuming Independence 74Don't Fall for the Gambler's Fallacy 74All Probabilities Are Conditional 75Don't Swap Dependencies 76Bayes' Theorem 76Ensure the Probabilities Have Meaning 79Calibration 80Rare Events Can, and Do, Happen 80Chapter Summary 81Chapter 7 Challenge the Statistics 83Quick Lessons on Inference 83Give Yourself Some Wiggle Room 84More Data, More Evidence 84Challenge the Status Quo 85Evidence to the Contrary 86Balance Decision Errors 88The Process of Statistical Inference 89The Questions You Should Ask to Challenge the Statistics 90What Is the Context for These Statistics? 90What Is the Sample Size? 91What Are You Testing? 92What Is the Null Hypothesis? 92Assuming Equivalence 93What Is the Significance Level? 93How Many Tests Are You Doing? 94Can I See the Confidence Intervals? 95Is This Practically Significant? 96Are You Assuming Causality? 96Chapter Summary 97Part Three Understanding the Data Scientist's ToolboxChapter 8 Search for Hidden Groups 101Unsupervised Learning 102Dimensionality Reduction 102Creating Composite Features 103Principal Component Analysis 105Principal Components in Athletic Ability 105PCA Summary 108Potential Traps 109Clustering 110k-Means Clustering 111Clustering Retail Locations 111Potential Traps 113Chapter Summary 114Chapter 9 Understand the Regression Model 117Supervised Learning 117Linear Regression: What It Does 119Least Squares Regression: Not Just a Clever Name 120Linear Regression: What It Gives You 123Extending to Many Features 124Linear Regression: What Confusion It Causes 125Omitted Variables 125Multicollinearity 126Data Leakage 127Extrapolation Failures 128Many Relationships Aren't Linear 128Are You Explaining or Predicting? 128Regression Performance 130Other Regression Models 131Chapter Summary 131Chapter 10 Understand the Classification Model 133Introduction to Classification 133What You'll Learn 134Classification Problem Setup 135Logistic Regression 135Logistic Regression: So What? 138Decision Trees 139Ensemble Methods 142Random Forests 143Gradient Boosted Trees 143Interpretability of Ensemble Models 145Watch Out for Pitfalls 145Misapplication of the Problem 146Data Leakage 146Not Splitting Your Data 146Choosing the Right Decision Threshold 147Misunderstanding Accuracy 147Confusion Matrices 148Chapter Summary 150Chapter 11 Understand Text Analytics 151Expectations of Text Analytics 151How Text Becomes Numbers 153A Big Bag of Words 153N-Grams 157Word Embeddings 158Topic Modeling 160Text Classification 163Naïve Bayes 164Sentiment Analysis 166Practical Considerations When Working with Text 167Big Tech Has the Upper Hand 168Chapter Summary 169Chapter 12 Conceptualize Deep Learning 171Neural Networks 172How Are Neural Networks Like the Brain? 172A Simple Neural Network 173How a Neural Network Learns 174A Slightly More Complex Neural Network 175Applications of Deep Learning 178The Benefits of Deep Learning 179How Computers "See" Images 180Convolutional Neural Networks 182Deep Learning on Language and Sequences 183Deep Learning in Practice 185Do You Have Data? 185Is Your Data Structured? 186What Will the Network Look Like? 186Artificial Intelligence and You 187Big Tech Has the Upper Hand 188Ethics in Deep Learning 189Chapter Summary 190Part Four Ensuring SuccessChapter 13 Watch Out for Pitfalls 193Biases and Weird Phenomena in Data 194Survivorship Bias 194Regression to the Mean 195Simpson's Paradox 195Confirmation Bias 197Effort Bias (aka the "Sunk Cost Fallacy") 197Algorithmic Bias 198Uncategorized Bias 198The Big List of Pitfalls 199Statistical and Machine Learning Pitfalls 199Project Pitfalls 200Chapter Summary 202Chapter 14 Know the People and Personalities 203Seven Scenes of Communication Breakdowns 204The Postmortem 204Storytime 205The Telephone Game 206Into the Weeds 206The Reality Check 207The Takeover 207The Blowhard 208Data Personalities 208Data Enthusiasts 209Data Cynics 209Data Heads 209Chapter Summary 210Chapter 15 What's Next? 211Index 215

O autorze

ALEX J. GUTMAN, PhD, is a Data Scientist, Corporate Trainer, and Accredited Professional Statistician. His professional focus is on statistical and machine learning and he has extensive experience working as a Data Scientist for the Department of Defense and two Fortune 50 companies.JORDAN GOLDMEIER is a Data Scientist, author, speaker, and community leader. He is a seven-time recipient of the Microsoft Most Valuable Professional Award and he has taught analytics to members of the Pentagon and Fortune 500 companies.