Tuesday, February 17, 2026
Forecast data in Connected Sheets using BigQuery ML and TimesFM
We are introducing a new capability in Connected Sheets for BigQuery that allows users to generate data forecasts directly within Google Sheets using the power of BigQuery ML (BQML) and Google’s state-of-the-art TimesFM model. Users can now predict future sales, demand volume, or other key business metrics without needing to write SQL, use Python, or customize and train their own models.
This feature is designed for simplicity and speed. TimesFM’s powerful foundation model is pre-trained on billions of data points from real-world datasets, enabling business users to generate sophisticated predictions on their data immediately without the complex, time-consuming training pipelines typically associated with machine learning.
Key features include:
- Simple configuration: Forecasts can be created from any existing BigQuery dataset or custom query using a user-friendly configuration pane in the Sheets UIl.
- Customizable parameters: Users can adjust forecast parameters, such as the prediction horizon for how far into the future the user wants to predict and confidence intervals, or use our default options to get started.
- Granular analysis: The feature supports breakouts by any given data dimension, allowing users to run multiple time series forecasts simultaneously (e.g., forecasting sales broken down by region or product category).
- Visual insights: For single time series forecasts, Sheets automatically generates a helpful chart visualizing the forecast alongside historical data.
Getting started
- Admins: There is no admin control for this feature.
- End users: Create a Connected Sheet. On the Preview view, click on “Advanced Analytics” and then “Create a Forecast”.
Rollout pace
- Rapid Release and Scheduled Release domains: Gradual rollout (up to 15 days for feature visibility) started on February 16, 2026
Availability
- Available to all Google Workspace customers and users with personal Google accounts
























