Behavioral Data Analysis with R and Python: Customer-Driven Data for Real Business Results

 Copyright....6

Table of Contents....7

Preface....13

Who This Book Is For....14

Who This Book Is Not For....15

R and Python Code....16

Code Environments....16

Code Conventions....17

Functional-Style Programming 101....18

Using Code Examples....18

Navigating This Book....19

Conventions Used in This Book....20

O’Reilly Online Learning....21

How to Contact Us....21

Acknowledgments....22

Part I. Understanding Behaviors....23

Chapter 1. The Causal-Behavioral Framework for Data Analysis....25

Why We Need Causal Analytics to Explain Human Behavior....26

The Different Types of Analytics....26

Human Beings Are Complicated....27

Confound It! The Hidden Dangers of Letting Regression Sort It Out....30

Data....31

Why Correlation Is Not Causation: A Confounder in Action....31

Too Many Variables Can Spoil the Broth....33

Conclusion....39

Chapter 2. Understanding Behavioral Data....41

A Basic Model of Human Behavior....42

Personal Characteristics....43

Cognition and Emotions....45

Intentions....46

Actions....47

Business Behaviors....48

How to Connect Behaviors and Data....50

Develop a Behavioral Integrity Mindset....50

Distrust and Verify....51

Identify the Category....52

Refine Behavioral Variables....54

Understand the Context....55

Conclusion....57

Part II. Causal Diagrams and Deconfounding....59

Chapter 3. Introduction to Causal Diagrams....61

Causal Diagrams and the Causal-Behavioral Framework....62

Causal Diagrams Represent Behaviors....63

Causal Diagrams Represent Data....64

Fundamental Structures of Causal Diagrams....68

Chains....68

Forks....72

Colliders....74

Common Transformations of Causal Diagrams....75

Slicing/Disaggregating Variables....76

Aggregating Variables....77

What About Cycles?....79

Paths....82

Conclusion....83

Chapter 4. Building Causal Diagrams from Scratch....85

Business Problem and Data Setup....86

Data and Packages....86

Understanding the Relationship of Interest....87

Identify Candidate Variables to Include....89

Actions....91

Intentions....92

Cognition and Emotions....93

Personal Characteristics....94

Business Behaviors....97

Time Trends....97

Validate Observable Variables to Include Based on Data....99

Relationships Between Numeric Variables....100

Relationships Between Categorical Variables....103

Relationships Between Numeric and Categorical Variables....106

Expand Causal Diagram Iteratively....108

Identify Proxies for Unobserved Variables....108

Identify Further Causes....109

Iterate....110

Simplify Causal Diagram....110

Conclusion....112

Chapter 5. Using Causal Diagrams to Deconfound Data Analyses....113

Business Problem: Ice Cream and Bottled Water Sales....114

The Disjunctive Cause Criterion....116

Definition....116

First Block....117

Second Block....119

The Backdoor Criterion....119

Definitions....120

First Block....122

Second Block....123

Conclusion....125

Part III. Robust Data Analysis....127

Chapter 6. Handling Missing Data....129

Data and Packages....131

Visualizing Missing Data....132

Amount of Missing Data....135

Correlation of Missingness....137

Diagnosing Missing Data....143

Causes of Missingness: Rubin’s Classification....146

Diagnosing MCAR Variables....148

Diagnosing MAR Variables....150

Diagnosing MNAR Variables....152

Missingness as a Spectrum....154

Handling Missing Data....158

Introduction to Multiple Imputation (MI)....159

Default Imputation Method: Predictive Mean Matching....162

From PMM to Normal Imputation (R Only)....163

Adding Auxiliary Variables....165

Scaling Up the Number of Imputed Data Sets....167

Conclusion....168

Chapter 7. Measuring Uncertainty with the Bootstrap....169

Intro to the Bootstrap: “Polling” Oneself Up....170

Packages....170

The Business Problem: Small Data with an Outlier....170

Bootstrap Confidence Interval for the Sample Mean....172

Bootstrap Confidence Intervals for Ad Hoc Statistics....177

The Bootstrap for Regression Analysis....179

When to Use the Bootstrap....182

Conditions for the Traditional Central Estimate to Be Sufficient....183

Conditions for the Traditional CI to Be Sufficient....183

Determining the Number of Bootstrap Samples....186

Optimizing the Bootstrap in R and Python....187

R: The boot Package....187

Python Optimization....190

Conclusion....191

Part IV. Designing and Analyzing Experiments....193

Chapter 8. Experimental Design: The Basics....195

Planning the Experiment: Theory of Change....196

Business Goal and Target Metric....197

Intervention....199

Behavioral Logic....201

Data and Packages....203

Determining Random Assignment and Sample Size/Power....204

Random Assignment....204

Sample Size and Power Analysis....207

Analyzing and Interpreting Experimental Results....221

Conclusion....224

Chapter 9. Stratified Randomization....225

Planning the Experiment....227

Business Goal and Target Metric....227

Definition of the Intervention....229

Behavioral Logic....230

Data and Packages....230

Determining Random Assignment and Sample Size/Power....231

Random Assignment....231

Power Analysis with Bootstrap Simulations....239

Analyzing and Interpreting Experimental Results....246

Intention-to-Treat Estimate for Encouragement Intervention....247

Complier Average Causal Estimate for Mandatory Intervention....248

Conclusion....254

Chapter 10. Cluster Randomization and Hierarchical Modeling....255

Planning the Experiment....256

Business Goal and Target Metric....256

Definition of the Intervention....256

Behavioral Logic....258

Data and Packages....258

Introduction to Hierarchical Modeling....259

R Code....260

Python Code....262

Determining Random Assignment and Sample Size/Power....264

Random Assignment....264

Power Analysis....266

Analyzing the Experiment....274

Conclusion....274

Part V. Advanced Tools in Behavioral Data Analysis....277

Chapter 11. Introduction to Moderation....279

Data and Packages....280

Behavioral Varieties of Moderation....280

Segmentation....281

Interactions....287

Nonlinearities....288

How to Apply Moderation....291

When to Look for Moderation?....292

Multiple Moderators....303

Validating Moderation with Bootstrap....309

Interpreting Individual Coefficients....311

Conclusion....317

Chapter 12. Mediation and Instrumental Variables....319

Mediation....320

Understanding Causal Mechanisms....320

Causal Biases....321

Identifying Mediation....323

Measuring Mediation....324

Instrumental Variables....329

Data....329

Packages....330

Understanding and Applying IVs....330

Measurement....333

Applying IVs: Frequently Asked Questions....336

Conclusion....337

Bibliography....339

Index....343

About the Author....359

Colophon....359

Harness the full power of the behavioral data in your company by learning tools specifically designed for behavioral data analysis. Common data science algorithms and predictive analytics tools treat customer behavioral data, such as clicks on a website or purchases in a supermarket, the same as any other data. Instead, this practical guide introduces powerful methods specifically tailored for behavioral data analysis.

Advanced experimental design helps you get the most out of your A/B tests, while causal diagrams allow you to tease out the causes of behaviors even when you can't run experiments. Written in an accessible style for data scientists, business analysts, and behavioral scientists, this practical book provides complete examples and exercises in R and Python to help you gain more insight from your data--immediately.

• Understand the specifics of behavioral data
• Explore the differences between measurement and prediction
• Learn how to clean and prepare behavioral data
• Design and analyze experiments to drive optimal business decisions
• Use behavioral data to understand and measure cause and effect
• Segment customers in a transparent and insightful way