ConversionFlow: A Library for Conversion Goals Analysis

The Conversion Goals Analysis project employs a two-stage model pipeline that integrates Bayesian Network modeling with an Optimization stage to derive optimal budget allocations. The Estimation stage builds a detailed probabilistic model to quantify customer journey dynamics via data loading, preprocessing, and MCMC sampling, while the Optimization stage utilizes a Genetic Algorithm to maximize conversion value by assigning budget weights that account for diminishing returns, constraints, and inter-node dependencies.

Contents:

• Comprehensive Guide to ConversionFlow
  • 1. Introduction
  • 2. Setup
  • 3. Running the Pipeline
  • 4. Interpreting the Results
  • 5. Stage 2: Running the Optimization
  • 6. Conclusion
• Getting Started with ConversionFlow
  • Installation
  • Basic Usage
  • Building Your First Model
• How-to Guides
  • How to Define a New Model Structure
• Explanation
  • Why Bayesian Networks?
  • The Two-Stage Pipeline
  • A Guide to Prior Selection
  • Estimation Process
  • Optimisation Process
  • Bayesian Networks in ConversionFlow
  • Causal Chain Analysis
  • Model Complexity Considerations
• Reference
  • API Reference
  • Configuration File Reference
  • Data Schema Reference
  • Project Architecture
• Troubleshooting
  • Frequently Asked Questions (FAQ)
• Testing
  • Testing Guide