What is BigQuery ML?
BigQuery ML enables training and running machine learning models directly in BigQuery using SQL syntax. Data analysts can build models without exporting data or learning a separate programming language. Models train on BigQuery data and deliver predictions as SQL queries.
For more complex model types such as deep neural networks, boosted trees, or AutoML, BigQuery ML relies on the Gemini Enterprise Agent Platform (the former Vertex AI service). Models can also be managed in the Model Registry and deployed to endpoints for online predictions.
Core Features
- SQL-based Training: CREATE MODEL with SQL syntax for model training
- Integrated Predictions: ML.PREDICT for batch and real-time predictions
- Automatic Feature Engineering: Automatic transformation of input data
- Model Registry: Versioning and management of trained models
- Platform Integration: Export and deploy models via the Gemini Enterprise Agent Platform for extended options
Typical Use Cases
Churn Prediction
SQL analysts create churn models on customer data. Training happens with CREATE MODEL, predictions with ML.PREDICT. No deep data science expertise required.
Demand Forecasting
ARIMA_PLUS and TimesFM models forecast time series like revenue or demand. Models automatically detect seasonality and trends in historical data.
Recommendation Systems
Matrix factorization creates recommendations from user-item interactions. Product recommendations and content personalization directly on BigQuery data.
Benefits
- No data export or ETL effort
- SQL knowledge sufficient for many common models
- Scales automatically with BigQuery infrastructure
- Seamless integration with BI tools and the Gemini Enterprise Agent Platform
Integration with innFactory
As a certified Google Cloud partner, innFactory supports you with BigQuery ML: use case identification, model design, feature engineering, and integration of ML predictions into business processes. We help with the decision between BigQuery ML and the Gemini Enterprise Agent Platform.
Available Tiers & Options
BigQuery ML
- No data export needed
- SQL-based
- Automatic feature engineering
- Limited model types for internal training
- Less control than dedicated ML platforms
Typical Use Cases
Technical Specifications
Frequently Asked Questions
What is BigQuery ML?
BigQuery ML enables training machine learning models directly in BigQuery using SQL. Data does not need to be exported, and SQL knowledge is sufficient for many common model types.
Which model types are supported?
Models trainable internally include linear and logistic regression, K-means clustering, matrix factorization, PCA, and time series models such as ARIMA_PLUS and TimesFM. More complex models like boosted trees, deep neural networks, or AutoML are trained via the Gemini Enterprise Agent Platform (formerly Vertex AI) and made usable within BigQuery ML.
How does BigQuery ML differ from the Gemini Enterprise Agent Platform?
BigQuery ML is optimized for SQL-based ML directly on BigQuery data. The Gemini Enterprise Agent Platform (formerly Vertex AI) offers more control, custom training, and MLOps features for more complex requirements; the two services are tightly integrated.
Can I use external models like TensorFlow or XGBoost?
Yes, BigQuery ML can import ONNX, TensorFlow, TensorFlow Lite, and XGBoost models, as well as reference remote models via Gemini Enterprise Agent Platform endpoints for predictions.
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of Google Cloud (official documentation). This page does not represent an offer by Google Cloud.
