What problem does it solve? Writing pandas-style dataframe and machine learning code against BigQuery requires knowing BigFrames-specific APIs, ordering modes, and ML packages that differ from standard pandas and scikit-learn, and mistakes cause slow queries, local memory overload, or runtime errors. ## Core Features & Use Cases - Dataframe API Best Practices: Generates BigFrames code that stays in the cloud, uses partial ordering mode, previews data with peek(), and avoids local materialization via to_pandas(). - BigQuery ML Integration: Trains and evaluates models with bigframes.bigquery.ml (linear regression, logistic regression) so training runs on BigQuery's engine instead of local memory. - Legacy BigFrames ML Guidance: Handles bigframes.ml specifics such as DataFrame-returning predict(), no random_state, automatic scaling, ARIMA Plus forecasting, and model persistence with to_gbq(). - Use Case: Ask for a logistic regression model on the public penguins dataset and receive complete BigFrames code that loads data with read_gbq, trains via bbq.ml.create_model, and evaluates without downloading data locally. ## Quick Start Write BigFrames code that trains a linear regression model on the BigQuery public penguins dataset to predict penguin weight.
