Interactive Notebooks¶

Get hands-on experience with OnlineRake through our comprehensive Jupyter notebooks!

These interactive notebooks provide complete tutorials with visualizations, real-time monitoring, and comprehensive examples that demonstrate the power of streaming weight calibration.

🚀 Getting Started¶

Perfect for newcomers to OnlineRake! Learn the basics with clear examples and visual proof that the algorithms work.

What you’ll learn:

  • Basic OnlineRake usage with SGD and MWU algorithms

  • Correcting feature bias in real-time survey data

  • Handling time-varying patterns in streaming data

  • Visual validation with comprehensive plots

  • Clear before/after comparisons showing success

⚡ Performance Comparison¶

Deep dive into algorithm performance across different bias scenarios.

What you’ll learn:

  • Comprehensive SGD vs MWU comparison

  • Testing across multiple bias patterns (linear, sudden, oscillating)

  • Performance metrics and statistical analysis

  • Algorithm selection guidance

  • Parameter tuning insights

🔬 Advanced Diagnostics¶

Master the monitoring and diagnostic capabilities for production deployments.

What you’ll learn:

  • Automatic convergence detection

  • Oscillation monitoring and problem diagnosis

  • Weight distribution evolution analysis

  • Real-time performance tracking

  • Production monitoring best practices

🎯 Quick Start Guide¶

  1. Install dependencies: pip install onlinerake[docs]

  2. Start with Getting Started: Master the basics first

  3. Compare algorithms: Understand when to use SGD vs MWU

  4. Learn diagnostics: Essential for production deployments

💡 Tips for Success¶

  • Run notebooks locally for the best interactive experience

  • Experiment with parameters to see their effects

  • Try your own data after completing the tutorials

  • Check diagnostics regularly in production environments

Each notebook is self-contained and includes all necessary imports and setup code. The visualizations clearly demonstrate that OnlineRake successfully corrects bias in streaming data!