APIs and reference

Access control

• Access control with IAM

  Access control reference for BigQuery ML including IAM roles and permissions.

Client libraries

• BigQuery client libraries

  Get started with BigQuery in your language of choice.

• `pandas-gbq` to BigQuery Python client library migration guide

  Guide for migrating code from `pandas-gbq` to the Python client library, `google-cloud-bigquery`.

• Migrating from the `datalab` Python package

  Guide for migrating code from the `datalab` Python package to the BigQuery Python client library.

Rest API

• REST API reference

  REST API reference for version 2 of the BigQuery API.

bq command-line tool reference

• `bq` command-line tool reference

  Reference for the `bq` command-line tool.

BigQuery ML standard SQL reference

• End-to-end user journey for each model

  An overview of all supported model types with their supported SQL statements and functions.

• The CREATE MODEL statement

  Creating and training models using the CREATE MODEL statement.

• The DROP MODEL statement

  Deleting models using the DROP MODEL statement.

• The EXPORT MODEL statement

  Exporting models using the EXPORT MODEL statement.

• The feature preprocessing functions

  Using the preprocessing functions for feature engineering.

• The ML.EVALUATE function

  Using the ML.EVALUATE function to evaluate models.

• The ML.ROC_CURVE function

  Using the ML.ROC_CURVE function to retrieve information about classification models.

• The ML.CONFUSION_MATRIX function

  Using the ML.CONFUSION_MATRIX function to return a confusion matrix for a given classification model and input data.

• The ML.PREDICT function

  Using the ML.PREDICT function to predict outcomes.

• The ML.RECOMMEND function

  Using the ML.RECOMMEND function to make recommendations.

• The ML.FORECAST function

  Using the ML.FORECAST function to perform time series forecasting.

• The ML.EXPLAIN_FORECAST function

  Using the ML.EXPLAIN_FORECAST function to perform time series forecasting with detailed components of the time series.

• The ML.ARIMA_EVALUATE function

  Using the ML.ARIMA_EVALUATE function to inspect the ARIMA_PLUS model evaluation metrics and possible error message.

• The ML.ARIMA_COEFFICIENTS function

  Using the ML.ARIMA_COEFFICIENTS function to inspect the ARIMA_PLUS model coefficients.

• The ML.TRAINING_INFO function

  Using the ML.TRAINING_INFO function to retrieve model training information.

• The ML.FEATURE_INFO function

  Using the ML.FEATURE_INFO function to retrieve model feature information.

• The ML.WEIGHTS function

  Using the ML.WEIGHTS function to see the underlying weights used by a model during prediction.

• The ML.EXPLAIN_PREDICT function

  Using the ML.EXPLAIN_PREDICT function to understand a model's predictions.

• The ML.DETECT_ANOMALIES function

  Using the ML.DETECT_ANOMALIES function to detect anomalies in the training data or in the prediction data.

• The ML.RECONSTRUCTION_LOSS function

  Using the ML.RECONSTRUCTION_LOSS function to compute reconstruction losses and detect anomalies.