/analytics-componentized-patterns

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

License

Analytics Componentized Patterns

From sample dataset to activation, these componentized patterns are designed to help you get the most out of BigQuery ML and other Google Cloud products in production.

Retail use cases

  • Recommendation systems:
    • How to build an end to end recommendation system with CI/CD MLOps pipeline on hotel data using BigQuery ML. (Code | Blogpost)
    • How to build a recommendation system on e-commerce data using BigQuery ML. (Code | Blogpost | Video)
    • How to build an item-item real-time recommendation system on song playlists data using BigQuery ML. (Code | Reference Guide)
  • Propensity to purchase model:
    • How to build an end-to-end propensity to purchase solution using BigQuery ML and Kubeflow Pipelines. (Code | Blogpost)
  • Activate on Lifetime Value predictions:
    • How to predict the monetary value of your customers and extract emails of the top customers to use in Adwords for example to create similar audiences. Automation is done by a combination of BigQuery Scripting, Stored Procedure and bash script. (Code)
  • Clustering:
    • How to build customer segmentation through k-means clustering using BigQuery ML. (Code | Blogpost)
  • Demand Forecasting:
    • How to build a time series demand forecasting model using BigQuery ML (Code | Blogpost | Video)

Gaming use cases

  • Propensity to churn model:
    • Churn prediction for game developers using Google Analytics 4 (GA4) and BigQuery ML. (Code | Blogpost | Video)

Financial use cases

  • Fraud detection
    • How to build a real-time credit card fraud detection solution. (Code | Blogpost | Video)

Questions? Feedback?

If you have any questions or feedback, please open up a new issue.

Disclaimer

This is not an officially supported Google product.

All files in this repository are under the Apache License, Version 2.0 unless noted otherwise.