Question

A retail analytics team already stores several years of sales history in BigQuery. Data analysts who are proficient in SQL want to build and run a product-demand forecasting model, but the company has no dedicated machine learning engineers and wants to avoid moving the data to a separate ML platform. Which Google Cloud capability best meets these needs?

Accepted Answer

Use BigQuery ML to create and train forecasting models directly in BigQuery using SQL statements

Answer

Export the data to Cloud Storage and hand it to a data science team to build models in a separate notebook environment

Answer

Load the sales data into Cloud SQL and write stored procedures to calculate forecasts

Answer

Stream the historical data through Dataflow to generate predictions in real time