Analytics Stories: Using Data to Make Good Things Happen

by Winston, Wayne L.

• ISBN13: 9781119646037
• ISBN10: 1119646030
• Edition: 1st
• Format: Paperback
• Copyright: 2020-09-16
• Publisher: Wiley

Summary

Inform your own analyses by seeing how one of the best data analysts in the world approaches analytics problems 

Analytics Stories: How to Make Good Things Happen is a thoughtful, incisive, and entertaining exploration of the application of analytics to real-world problems and situations. Covering fields as diverse as sports, finance, politics, healthcare, and business, Analytics Stories bridges the gap between the oft inscrutable world of data analytics and the concrete problems it solves. 

Distinguished professor and author Wayne L. Winston answers questions like: 

• Was Liverpool over Barcelona the greatest upset in sports history? 
• Was Derek Jeter a great infielder 
• What's wrong with the NFL QB rating? 
• How did Madoff keep his fund going? 
• Does a mutual fund’s past performance predict future performance? 
• What caused the Crash of 2008? 
• Can we predict where crimes are likely to occur? 
• Is the lot of the American worker improving? 
• How can analytics save the US Republic? 
• The birth of evidence-based medicine: How did James Lind know citrus fruits cured scurvy? 
• How can I objectively compare hospitals? 
• How can we predict heart attacks in real time? 
• How does a retail store know if you're pregnant? 
• How can I use A/B testing to improve sales from my website? 
• How can analytics help me write a hit song? 

Perfect for anyone with the word “analyst” in their job title, Analytics Stories illuminates the process of applying analytic principles to practical problems and highlights the potential pitfalls that await careless analysts.  

Author Biography

Wayne L. Winston is Professor Emeritus of Decision Sciences at the Kelley School of Business at Indiana University. He currently teaches Sports Analytics at IU, and has taught analytics to organizations including Microsoft, eBay, Cisco, Deloitte, the U.S. military, Eli Lilly, JP Morgan, and more. A two-time Jeopardy! Champion, he has consulted on analytics for two top NBA teams.

Table of Contents

Introduction xxvii

Part I What Happened? 1

Chapter 1 Preliminaries 3

Basic Concepts in Data Analysis 3

What Is a Random Variable? 9

Excel Calculations 13

Chapter 2 Was the 1969 Draft Lottery Fair? 17

The Data 17

The Analysis 18

Excel Calculations 20

Chapter 3 Who Won the 2000 Election: Bush or Gore? 23

Projecting the Undervotes 24

What Happened with the Overvotes? 25

The Butterfl y Did It! 25

Excel Calculations 28

Chapter 4 Was Liverpool Over Barcelona the Greatest Upset in Sports History? 31

How Should We Rank Upsets? 31

Leicester Wins the 2015–2016 Premier League 32

#16 Seed UMBC Beats #1 Seed Virginia 33

The Jets Win Super Bowl III 33

Other Big Upsets 34

Chapter 5 How Did Bernie Madoff Keep His Fund Going? 35

The Mathematics of Ponzi Schemes 36

Madoff’s Purported Strategy 37

The Sharpe Ratio Proves Madoff Was a Fraud 39

Benford’s Law and Madoff’s Fraud 40

Excel Calculations 41

Chapter 6 Is the Lot of the American Worker Improving? 45

Is U.S. Family Income Skewed? 45

Median Income and Politics 46

Causes of Increasing U.S. Income Inequality 48

Money Isn’t Everything: The Human

Development Index 50

Create Your Own Ranking of Well-Being 50

Are Other Countries Catching Up to the U.S.? 51

Excel Calculations 52

Chapter 7 Measuring Income Inequality with the Gini, Palm, and Atkinson Indices 53

The Gini Index 53

The Palma Index 56

The Atkinson Index 57

Excel Calculations 59

Chapter 8 Modeling Relationships Between Two Variables 61

Examples of Relationships Between Two Variables 61

Finding the Best-Fitting (Least Squares) Line 62

Computing the Beta of a Stock 63

What Is a Good R2? 64

Correlation and R2 65

We are Not Living in a Linear World 67

Excel Calculations 69

Chapter 9 Intergenerational Mobility 73

Absolute Intergenerational Mobility 74

Intergenerational Elasticity 74

Rank-Rank Mobility 75

Comparing IGE and Rank-Rank Mobility 75

Measuring Mobility with Quintiles 78

The Great Gatsby Curve 80

Excel Calculations 82

Chapter 10 Is Anderson Elementary School a Bad School? 85

How Can We Adjust for Family Income? 86

Estimating the Least Squares Line 86

Can We Compare Standardized Test Performance for Students in Different States? 86

Excel Calculations 87

Chapter 11 Value-Added Assessments of Teacher Effectiveness 89

Simple Gain Score Assessment 90

Covariate Adjustment Assessment 91

Layered Assessment Model 91

Cross-Classified Constant Growth Assessment 91

Problems with VAA 93

How Much Is a Good Teacher Worth? 94

Excel Calculations 95

Chapter 12 Berkeley, Buses, Cars, and Planes 97

Simpson’s Paradox and College Admissions 98

The Waiting Time Paradox 100

When Is the Average of 40 and 80 Not 60? 100

Why Pre COVID Were There Never Empty

Seats on My Flight? 101

Excel Calculations 101

Chapter 13 Is Carmelo Anthony a Hall of Famer? 103

What Metric Defines Basketball Ability? 104

Wins Above Replacement Player (WARP) 105

Manu, Melo, Dirk, and Dwayne 106

How Do 25,000 Points Lead to So Few Wins? 106

Chapter 14 Was Derek Jeter a Great Fielder? 109

Fielding Statistics: The First Hundred Years 109

Range Factor 110

The Fielding Bible: A Great Leap Forward 111

The Next Frontier 112

Chapter 15 “Drive for Show and Putt for Dough?” 115

Strokes Gained 115

The Myth Exposed 116

Chapter 16 What’s Wrong with the NFL QB Rating? 117

NFL Quarterback Rating 117

ESPN’s Total Quarterback Rating 124

Excel Calculations 125

Chapter 17 Some Sports Have All the Luck 127

Skill vs. Luck: The Key Idea 127

The Results 129

Chapter 18 Gerrymandering 131

A Stylized Example 132

The Mathematics of Gerrymandering 136

Chapter 19 Evidence-Based Medicine 143

James Lind and Scurvy: The Birth of Evidence-Based Medicine 143

The Randomized Streptomycin Tuberculosis Trial 145

Excel Calculations 146

Hormone Replacement: Good or Bad? 148

Chapter 20 How Do We Compare Hospitals? 151

Ratings Criteria 152

Conclusion 156

Excel Calculations 157

Chapter 21 What Is the Worst Health Care Problem in My Country? 159

Disability-Adjusted Life Years 159

Determination of Disability Weights 160

To Age Weight or Discount, That Is the Question 162

Key Facts About World Health 163

Part II What Will Happen? 167

Chapter 22 Does a Mutual Fund’s Past Performance Predict Future Performance? 169

Mutual Fund Basics 170

Morningstar Ratings 170

Risk-Adjusting Fund Returns 171

How Well Do Morningstar Star Ratings

Predict a Fund’s Future Performance? 175

The Effect of Expense Ratio on Long-Term Performance 177

Excel Calculations 178

Chapter 23 Is Vegas Good at Picking NFL Games? 181

How NFL Betting Works 181

Bias and Accuracy 184

Vegas Forecasts are Unbiased 185

Totals Predictions and Money Line Predictions are Unbiased 188

NFL Accuracy: The Line vs. the Computers 188

A System Works Until It Doesn’t 189

Chapter 24 Will My New Hires Be Good Employees? 191

What Data Do We Need to Determine Attributes That Best Predict Employee Performance? 192

Besides GMA, Not Much Affects Job Performance 196

Excel Calculations 197

Chapter 25 Should I Go to State U or Princeton? 199

Analyzing Princeton vs. Penn State 200

Excel Calculations 202

Chapter 26 Will My Favorite Sports Team Be Great Next Year? 203

Francis Galton and Regression to the Mean 203

Regression to the Mean in the NFL and the NBA 204

Excel Calculations 207

Chapter 27 How Did Central Bankers Fail to Predict the 2008 Recession? 209

The Inverted Yield Curve 210

The Sahm Rule: Early Warning Signal for Recession 211

Control Charts and the Housing Price/Rent Ratio 211

Excel Calculations 215

Chapter 28 How Does Target Know If You’re Pregnant? 219

What Available Data Can Be Used

to Identify Pregnant Women? 220

Problems Arise 220

An Example of a Pregnancy Prediction Score 221

Chapter 29 How Does Netflix Recommend Movies and TV Shows? 225

User-Based Collaborative Filtering 226

Item-Based Filtering 229

Chapter 30 Can We Predict Heart Attacks in Real Time? 233

Posterior Probability 234

Sensitivity and Specifi city 235

ROC Curve 235

Back to the Apple Heart Study 237

AliveCor and KardiaBand 239

Chapter 31 Is Proactive Policing Effective? 241

Hot Spots Policing 242

Predictive Policing 243

CCTV 244

Stop and Frisk 244

Broken Windows 246

Excel Calculations 247

Chapter 32 Guess How Many are Coming to Dinner? 249

Which Parameters Must Be Estimated? 250

The Data 252

The Results 253

Which Factor Really Matters? 254

Excel Calculations 254

Chapter 33 Can Prediction Markets Predict the Future? 259

Examples of Trade Contracts 260

Prediction Market Trading Mechanisms 261

Accuracy of Prediction Markets and Wisdom of Crowds 262

Chapter 34 The ABCs of Polling 265

Why are 1,112 People Enough to Represent U.S. Voters? 265

Why Doesn’t a Larger Population Require a Larger Sample Size? 267

So, What Can Go Wrong? 268

Rating Polls 271

Chapter 35 How Did Buzzfeed Make the Dress Go Viral? 273

Measuring Instagram Engagement 274

Tweets Do Not Always Go Viral Immediately 274

Do the First Few Days Predict the Future of a Meme? 275

Chapter 36 Predicting Game of Thrones TV Ratings 277

What Does Google Trends Tell Us? 277

Predicting the Present with Google Trends 278

Using Google Trends to Forecast GOT Ratings 279

Excel Calculations 281

Part III Why Did It Happened? 283

Chapter 37 Does Smoking Cause Lung Cancer? 285

Correlation and Causation Redux 285

The Key Evidence 286

Could Air Pollution Have Caused Lung Cancer? 287

The Cigarette Companies Hit Back 287

Excel Calculations 288

Chapter 38 Why are the Houston Rockets a Good Basketball Team? 291

NBA Shooting Math 101 292

Zach LaVine Battles the Bulls’ Analytics Department 295

Conclusion 296

Excel Calculations 296

Chapter 39 Why Have Sacrifice Bunts and Intentional Walks Nearly Disappeared? 297

The Case Against Bunting 298

Bunting Against the Shift 299

Why are Intentional Walks on the Decline? 300

Chapter 40 Do NFL Teams Pass Too Much and Go for It Often Enough on Fourth Down? 301

The Ascent of Passing 301

Fourth Down Strategy 303

New Data Partially Vindicates the Coaches 304

Teams Should Go for Two More Often 306

Chapter 41 What Caused the 1854 London Cholera Outbreak? 307

Cholera 307

Snow and the Broad Street Pump 308

Snow’s Randomized Controlled Trial 310

Conclusion 311

Excel Calculations 312

Chapter 42 What Affects the Sales of a Retail Product? 313

Painter’s Tape 313

Estimating the Model Parameters 315

Excel Calculations 316

Chapter 43 Why Does the Pareto Principle Explain So Many Things? 319

Power Laws 320

Why Do Incomes Follow the Pareto Principle? 322

Why Do a Few Websites Get Most of the Hits? 323

Excel Calculations 324

Chapter 44 Does Where You Grow Up Matter? 327

Quasi-Experimental Design vs. Randomized Controlled Trials 328

What Drives Neighborhood Differences in Upward Mobility? 329

How Can We Make Things Better? 330

Chapter 45 The Waiting is the Hardest Part 333

Which Factors Influence the Performance of a Queueing System? 334

Operating Characteristics of a Queueing System 334

How Does Variability Degrade the Performance of a Queueing System? 335

Calculating the Operating Characteristics of a Queueing System 336

Excel Calculations 338

Chapter 46 are Roundabouts a Good Idea? 339

What Is a Roundabout? 340

History of Roundabouts 340

Benefi ts of Roundabouts 341

Disadvantages of Roundabouts 343

Roundabout Capacity 344

Roundabouts and Revolutions 345

Chapter 47 Red Light, Green Light, or No Light? 347

What Causes Traffic Jams? 347

How Should We Set the Lights? 348

Ramp Meters and Equity 349

Measuring the Impact of Ramp Meters 350

The Twin Cities Metering Holiday 350

Part IV How Do I Make Good Things Happen? 351

Chapter 48 How Can We Improve K–12 Education? 353

Tennessee’s STAR Study on K–2 Class Size 355

Cost–Benefi t Analysis 356

Can Predictive Analytics Increase Enrollment and Performance in Eighth-Grade Algebra I? 360

Excel Calculations 360

Chapter 49 Can A/B Testing Improve My Website’s Performance? 363

Improving Obama’s Fundraising in 2008 364

The Mechanics of Resampling 365

Excel Calculations 366

Chapter 50 How Should I Allocate My Retirement Portfolio? 369

The Basic Portfolio Optimization Model 369

The Effi cient Frontier 372

Diffi culties in Implementing the Markowitz Model 374

Excel Calculations 374

Chapter 51 How Do Hedge Funds Work? 377

Growth in Hedge Funds and Hedge Fund Fee Structure 378

Shorting a Stock 378

Long/Short and Market-Neutral Strategies 378

Convertible Arbitrage 380

Merger Arbitrage 382

Global Macro Strategy 383

Hedge Fund Performance 384

The George Costanza Portfolio 384

Excel Calculations 385

Chapter 52 How Much Should We Order and When Should We Order? 389

The Economic Order Quantity Model 389

Reorder Points, Service Levels, and Safety Stock 392

Excel Calculations 393

Chapter 53 How Does the UPS Driver Know the Order to Deliver Packages? 397

Why Is the Traveling Salesperson Problem So Hard? 398

Solving the Traveling Salesperson Problem 399

The Traveling Salesperson Problem in the Real World 400

Excel Calculations 401

Chapter 54 Can Data Win a Presidential Election? 405

Democratic Presidential Analytics 405

The GOP Strikes Back 409

Cambridge Analytica and the 2016 Election 411

Excel Calculations 412

Chapter 55 Can Analytics Save Our Republic? 415

Arrow’s Impossibility Theorem 416

It’s Not Easy to Pick a Winner! 417

Ranked-Choice Voting 419

Approval Voting 420

Quadratic Voting 420

Excel Calculations 421

Chapter 56 Why Do I Pay Too Much on eBay? 423

How Many Pennies in the Jar? 423

The Importance of Asymmetric Information 424

The Winner’s Curse and Offshore Oil Leases 424

Sports Free Agents and the Winner’s Curse 425

Can You Avoid the Winner’s Curse? 425

Excel Calculations 427

Chapter 57 Can Analytics Recognize, Predict, or Write a Hit Song? 429

How Does Shazam Know What Song You are Listening To? 430

How Did Hit Song Science Know Norah Jones’s Album Would Be a Smash? 431

Can Artifi cial Intelligence Write a Good Song? 433

Chapter 58 Can an Algorithm Improve Parole Decisions? 437

An Example of Risk Scores 438

ProPublica Criticizes Risk Scores 441

Skeem and Lowenkamp and PCRA 443

Machine Learning and Parole Decisions 444

Chapter 59 How Do Baseball Teams Decide Where to Shift Fielders? 449

The Debut of the Shift 449

The Return of the Shift 450

Empirical Evidence on the Shift 452

Why Not Just Beat the Shift? 452

Excel Calculations 453

Chapter 60 Did Analytics Help the Mavericks Win the 2011 NBA Title? 457

How Can You Evaluate a Basketball Player? 457

From Player Ratings to Lineup Ratings 459

Chapter 61 Who Gets the House in the Hamptons? 463

The Basic Idea 464

What Asset Division Is Best? 465

Excel Calculations 466

Index 469