Analytics Data → Machine Learning

GCP • AI/ML • Advanced • Retail

Architecture Diagram

Overview

Retail businesses lose time and money when orders are returned.

Some returns are normal, but if teams can identify risky orders earlier, they can plan better follow-ups, improve customer experience, and reduce operational surprises.

In this pipeline, you will build a simple ML workflow on GCP.

You will create features from order, customer, item, and return data, train a return-risk model using BigQuery ML, score recent orders, and store prediction results for reporting or downstream use.

What You Will Build

  • Build a model-ready feature table from retail data
  • Create labels from matured return history
  • Train a return-risk model using BigQuery ML
  • Evaluate model performance
  • Score recent orders for return risk
  • Store prediction results in BigQuery
  • Track model runs, metrics, and prediction status
  • Orchestrate the ML workflow using Cloud Composer

Tech Stack

BigQuery • BigQuery ML • Cloud Composer Apache Airflow • SQL

Learning Outcomes

After completing this pipeline, you will be able to:

  1. Build feature tables from business data
  2. Create labels from historical outcomes
  3. Train a machine learning model using BigQuery ML
  4. Evaluate model performance using BigQuery ML
  5. Score recent orders using an ML model
  6. Store prediction results for reporting and downstream use
  7. Track model metrics and pipeline audit details
  8. Orchestrate ML workflows using Cloud Composer
View Courses & Pricing →