Project Synopsis

The project encompasses these main stages:

1. Web analytics data processing and preparation
2. Feature engineering to establish an article engagement metric
3. Training a recommendation algorithm using cloud-based ML tools
4. Applying the model to generate content suggestions

Repository Contents

• create_table.sql: SQL code for preparing and transforming raw analytics data
• train.sql: SQL code for building the recommendation algorithm
• predict.sql: SQL code for producing recommendations with the trained model
• bqml_ga360.ipynb: Jupyter notebook detailing the complete process with explanations

Key Aspects

• Uses session duration as an indicator of article interest
• Implements data scaling and normalization methods
• Leverages cloud-based matrix factorization capabilities
• Demonstrates large-scale data handling in a cloud environment

Quick Start Guide

1. Confirm access to a cloud-based dataset containing Google Analytics information
2. Execute the SQL scripts in this sequence:
   • create_table.sql
   • train.sql
   • predict.sql