Skip to main content
Get the results of analytics (regression, correlation, chi-square test and so on) in a published dashboard or note as structured data.

Two steps

1. Find the analytics name

Run fetch. In metadata.blocks of the result, the entries whose kind is analytics hold the name of the analytics and its analysis_type. In the next step, pass the name, not the title.
Replace <id> with an id from a search result. The result below was run on the dashboard Sales Orders Dashboard (id pQ9rS0tU1v). Long output is trimmed. Result
If the block with kind analytics shows a name, you have the name.

2. Get the result

Pass the name to --analytics and run analytics get-result.
Replace <id> with the id of the dashboard and <analytics> with the name of the analytics. The result below is for the analytics oP8qR9sT (Monthly Sales Forecast, do_prophet). Long output is trimmed. Result
The result holds the model definition (target, predictors, N), result tables with stable column ids, headline statistics split into training and test, the settings it ran with, and explanation texts, each marked as built from the numbers or written by AI. Values are as displayed, and coefficients are unstandardized. This example is a time series forecast, so target and predictors in model_definition are empty. --limit is the maximum number of rows per table (up to 100). There is no way to read on. By default it returns the tables that answer the analysis. To choose tables, pass ids from available_tables to --tables.
If result.tables holds result tables, you have the result.

Troubleshooting

Every flag is described in analytics get-result.