Challenge 7: Notebooks for data scientists

Challenge 7: Notebooks for data scientists

Previous Challenge Next Challenge

Introduction

BigQuery Studio and SQL are great tools for data analytics, but data scientists also enjoy working with interactive notebooks using Python. Luckily BigQuery provides Python notebooks capabilites integrated with BigQuery, making it possible to run Python and (serverless) Spark tasks.

Description

We’ve already designed a Python notebook for this challenge. Upload that to BigQuery, run the notebook interactively until you get to the cell for creating the model. Edit the cell and add the necessary SQL to create a BigQuery ML model, and run the notebook to completion.

Warning
Since notebooks on BigQuery use End User Credentials, make sure that you’re running this notebook as a user that doesn’t have the row level security filter applied to limit the data.

Success Criteria

• All the cells from the provided Python notebook has been run successfully.
• There’s a new Logistic Regression model, churn_model, in the dwh dataset that predicts whether a customer will churn or not, trained on the prepared training data.

Learning Resources

• BigQuery Python Notebooks
• Introduction to AI/ML in BigQuery
• Tutorial on building a BQML model

Previous Challenge Next Challenge