Preface

13

Who Is This Book For?

13

Analyst with Some Coding Experience

13

Undergraduate Student Looking for Applied Experience

14

Nontechnical Leader of an Analytics Team

15

The Structure of This Book

16

Conclusion

18

1 Introduction

19

1.1 Aligning on Nomenclature

19

1.2 Learning to Google (or Prompt)

21

1.2.1 What Can You Find with Google?

21

1.2.2 Prompting

25

1.3 Predictions for Generative AI’s Impact on Machine Learning

26

1.4 Summary

26

2 Getting Started

27

2.1 GitHub

27

2.1.1 Creating an Account

28

2.1.2 GitHub in This Book

30

2.2 Anaconda

30

2.2.1 Creating an Account

31

2.2.2 Creating Projects and Uploading Data

34

2.2.3 Anaconda in This Book

37

2.3 Summary

38

3 Introduction to Our Use Cases

39

3.1 Importance of Understanding the Business Problem

39

3.1.1 Business Reviews

40

3.1.2 Definition of Success

40

3.2 Use Case 1: The Retail Tyrant

41

3.2.1 Details of the Request

41

3.2.2 History of the Request

42

3.2.3 Relationship with the Stakeholder

42

3.2.4 Use Case Questions

43

3.2.5 Use Case Answers

44

3.3 Use Case 2: Customer Retention

47

3.3.1 Details of the Request

47

3.3.2 History of the Request

47

3.3.3 Relationship with the Stakeholder

48

3.3.4 Use Case Questions

48

3.3.5 Use Case Answers

49

3.4 Use Case 3: Crime Predictions

50

3.4.1 Details of the Request

51

3.4.2 History of the Request

51

3.4.3 Relationship with the Stakeholder

51

3.4.4 Use Case Questions

52

3.4.5 Use Case Answers

52

3.5 Summary

53

4 Starting with the Data

55

4.1 Types of Data Sources

55

4.1.1 Manual

56

4.1.2 Automated

59

4.1.3 Data Sources for Our Use Cases

60

4.2 Data Exploration

66

4.2.1 Data Types

66

4.2.2 Data Visualization

77

4.2.3 Descriptive Statistics

105

4.2.4 Correlation Analysis

114

4.3 Data Cleaning (For Now)

120

4.3.1 Why Isn’t Data Already Clean?

121

4.3.2 Overview of Cleaning for Regression Models

122

4.3.3 Inaccurate Data

123

4.3.4 Missing Data

123

4.3.5 Dummy Coding

131

4.3.6 Dimensionality Reduction

161

4.4 Summary

178

5 Picking Your Model

181

5.1 The Simpler the Model, the Better

181

5.2 Model Decision Framework

183

5.2.1 How Important Is Interpretability?

184

5.2.2 How Many Rows and Columns?

184

5.2.3 What Is Being Predicted?

185

5.3 Train-Test Split

187

5.4 Regression Models

189

5.4.1 What Are Regression Models?

189

5.4.2 Multicollinearity

192

5.4.3 Linear Regression

192

5.4.4 Logistic Regression

211

5.5 Machine Learning Models

221

5.5.1 Decision Tree

222

5.5.2 Random Forest

252

5.5.3 Gradient Boosting Machine

271

5.6 Clustering

291

5.6.1 What Is Clustering?

292

5.6.2 Picking the Number of Clusters

294

5.6.3 Behind the Scenes of Clustering

296

5.7 Summary

297

6 Evaluating the Model and Iterating

299

6.1 Importance of Picking Validation Metrics

299

6.2 Validation Metrics

301

6.2.1 Accuracy

302

6.2.2 Confusion Matrix

302

6.2.3 Precision

305

6.2.4 Recall

305

6.2.5 F1 Score

305

6.2.6 Area Under the Curve

306

6.2.7 R-Squared

307

6.2.8 Mean Squared Error

309

6.2.9 Mean Absolute Error

309

6.2.10 Metric Summary

309

6.3 K-Fold Cross-Validation

311

6.4 Business Validations

311

6.4.1 Legal Considerations

312

6.4.2 Ethical Considerations

313

6.5 Machine Learning Interpretability

314

6.5.1 Regression Models

314

6.5.2 Tree-Based Models

316

6.6 Iterating on the Model

321

6.6.1 Feature Engineering

322

6.6.2 Remove Variables

324

6.6.3 Add New Data

325

6.7 Application to Use Cases

328

6.7.1 Use Case 1

328

6.7.2 Use Case 2

348

6.7.3 Use Case 3

362

6.8 Summary

374

7 Implementing, Monitoring, and Measuring the Model

375

7.1 Implementing Your Model for Predictions

375

7.1.1 Don’t Train the Model Each Time

376

7.1.2 Predictions for Our Use Cases

376

7.1.3 Saving Your Predictions

392

7.1.4 Practical Approaches to Consider

393

7.2 Model Monitoring

394

7.2.1 Importance of Model Monitoring

394

7.2.2 What to Monitor

395

7.2.3 Considerations for Model Monitoring

399

7.2.4 Retraining the Model

400

7.3 Measuring the Impact of Your Model

401

7.3.1 Business Sniff Test

401

7.3.2 Experiments

403

7.4 Summary

426

8 Closing Thoughts

427

8.1 Learning How to Learn with Generative AI

427

8.2 Learning How to Learn with Use Cases

428

8.3 Explore and Visualize Your Data

428

8.4 Cleaning Your Data and Dummy Coding

429

8.5 Machine Learning Models

430

8.6 Hyperparameters and Grid Search

430

8.7 Variable Lagging

431

8.8 The End

431

8.9 Acknowledgments

431