AI-Powered Business Intelligence

Table of contents

1. Preface
   1. Who Should Read This Book
   2. Microsoft Power BI and Azure
   3. Learning Objectives
   4. Navigating This Book
   5. Conventions Used in This Book
   6. Using Code Examples
   7. O’Reilly Online Learning
   8. How to Contact Us
   9. Acknowledgments
2. 1. Creating Business Value with AI
   1. How AI Is Changing the BI Landscape
   2. Common AI Use Cases for BI
      1. Automation and Ease of Use
      2. Better Forecasting and Predictions
      3. Leveraging Unstructured Data
   3. Getting an Intuition for AI and Machine Learning
   4. Mapping AI Use Case Ideas to Business Impact
   5. Summary
3. 2. From BI to Decision Intelligence: Assessing Feasibility for AI Projects
   1. Putting Data First
   2. Assessing Data Readiness with the 4V Framework
      1. Combining 4Vs to Assess Data Readiness
   3. Choosing to Make or Buy AI Services
      1. AI as a Service
      2. Platform as a Service
      3. Infrastructure as a Service
      4. End-to-End Ownership
   4. Basic Architectures of AI Systems
      1. User Layer
      2. Data Layer
      3. Analysis Layer
   5. Ethical Considerations
   6. Creating a Prioritized Use Case Roadmap
      1. Mix Champions and Quick Wins
      2. Identify Common Data Sources
      3. Build a Compelling Vision
   7. Summary
4. 3. Machine Learning Fundamentals
   1. The Supervised Machine Learning Process
      1. Step 1: Collect Historical Data
      2. Step 2: Identify Features and Labels
      3. Step 3: Split Your Data into Training and Test Sets
      4. Step 4: Use Algorithms to Find the Best Model
      5. Step 5: Evaluate the Final Model
      6. Step 6: Deploy
      7. Step 7: Perform Maintenance
   2. Popular Machine Learning Algorithms
      1. Linear Regression
      2. Decision Trees
      3. Ensemble Learning Methods
   3. Deep Learning
      1. Natural Language Processing
      2. Computer Vision
      3. Reinforcement Learning
   4. Machine Learning Model Evaluation
      1. Evaluating Regression Models
      2. Evaluating Classification Models
      3. Evaluating Multiclassification Models
   5. Common Pitfalls of Machine Learning
      1. Pitfall 1: Using Machine Learning When You Don’t Need It
      2. Pitfall 2: Being Too Greedy
      3. Pitfall 3: Building Overly Complex Models
      4. Pitfall 4: Not Stopping When You Have Enough Data
      5. Pitfall 5: Falling for the Curse of Dimensionality
      6. Pitfall 6: Ignoring Outliers
      7. Pitfall 7: Taking Cloud Infrastructure for Granted
   6. Summary
5. 4. Prototyping
   1. What Is a Prototype, and Why Is It Important?
   2. Prototyping in Business Intelligence
   3. The AI Prototyping Toolkit for This Book
   4. Working with Microsoft Azure
      1. Sign Up for Microsoft Azure
      2. Create an Azure Machine Learning Studio Workspace
      3. Create an Azure Compute Resource
      4. Create Azure Blob Storage
   5. Working with Microsoft Power BI
   6. Summary
6. 5. AI-Powered Descriptive Analytics
   1. Use Case: Querying Data with Natural Language
      1. Problem Statement
      2. Solution Overview
      3. Power BI Walk-Through
   2. Use Case: Summarizing Data with Natural Language
      1. Problem Statement
      2. Solution Overview
      3. Power BI Walk-Through
   3. Summary
7. 6. AI-Powered Diagnostic Analytics
   1. Use Case: Automated Insights
      1. Problem Statement
      2. Solution Overview
      3. Power BI Walk-Through
   2. Summary
8. 7. AI-Powered Predictive Analytics
   1. Prerequisites
   2. About the Dataset
   3. Use Case: Automating Classification Tasks
      1. Problem Statement
      2. Solution Overview
      3. Model Training with Microsoft Azure Walk-Through
      4. What Is an AutoML Job?
      5. Evaluating the AutoML Outputs
      6. Model Deployment with Microsoft Azure Walk-Through
      7. Getting Model Predictions with Python or R
      8. Model Inference with Power BI Walk-Through
      9. Building the AI-Powered Dashboard in Power BI
   4. Use Case: Improving KPI Prediction
      1. Problem Statement
      2. Solution Overview
      3. Model Training with Microsoft Azure Walk-Through
      4. Model Deployment with Microsoft Azure Walk-Through
      5. Getting Model Predictions with Python or R
      6. Model Inference with Power BI Walk-Through
      7. Building the AI-Powered Dashboard in Power BI
   5. Use Case: Automating Anomaly Detection
      1. Problem Statement
      2. Solution Overview
      3. Enabling AI Service on Microsoft Azure Walk-Through
      4. Getting Model Predictions with Python or R
      5. Model Inference with Power BI Walk-Through
      6. Building the AI-Powered Dashboard in Power BI
   6. Summary
9. 8. AI-Powered Prescriptive Analytics
   1. Use Case: Next Best Action Recommendation
      1. Problem Statement
      2. Solution Overview
      3. Setting Up the AI Service
      4. How Reinforcement Learning Works with the Personalizer Service
      5. Setting Up Azure Notebooks
      6. Simulating User Interactions
      7. Running the Simulation with Python
      8. Evaluate Model Performance in Azure Portal
      9. Model Inference with Power BI Walk-Through
      10. Building the AI-Powered Dashboard in Power BI
   2. Summary
10. 9. Leveraging Unstructured Data with AI
    1. Use Case: Getting Insights from Text Data
       1. Problem Statement
       2. Solution Overview
       3. Setting Up the AI Service
       4. Setting Up the Data Pipeline
       5. Model Inference with Power BI Walk-Through
       6. Building the AI-Powered Dashboard in Power BI
    2. Use Case: Parsing Documents with AI
       1. Problem Statement
       2. Solution Overview
       3. Setting Up the AI Service
       4. Setting Up the Data Pipeline
       5. Model Inference with Power BI Walk-Through
       6. Building the AI-Powered Dashboard in Power BI
    3. Use Case: Counting Objects in Images
       1. Problem Statement
       2. Solution Overview
       3. Setting Up the AI Service
       4. Setting Up the Data Pipeline
       5. Model Inference with Power BI Walk-Through
       6. Building the AI-Powered Dashboard in Power BI
    4. Summary
11. 10. Bringing It All Together: Building an AI-Powered Customer Analytics Dashboard
    1. Problem Statement
    2. Solution Overview
    3. Preparing the Datasets
    4. Allocating a Compute Resource
    5. Building the ML Workflow
       1. Adding Sentiment Data to the Workflow
       2. Deploying the Workflow for Inference
    6. Building the AI-Powered Dashboard in Power BI
       1. Anomaly Detection
       2. Predictive Analytics
       3. AI-Powered Descriptive Analytics
       4. Unstructured Data
    7. Summary
12. 11. Taking the Next Steps: From Prototype to Production
    1. Discovery Versus Delivery
    2. Success Criteria for AI Product Delivery
       1. People
       2. Processes
       3. Data
       4. Technology
    3. MLOps
    4. Get Started by Delivering Complete Increments
    5. Conclusion
13. Index
14. About the Author