Getting Genie to Generate SPC (Control) Charts Reliably

curious_rabbit
New Contributor II

‎12-11-2025 10:55 PM

Hi everyone!

I’m working on getting Genie to accurately generate Statistical Process Control (SPC) charts when prompted.  
I'm looking for suggestions on how to best approach this. So far, I’ve tried using pre-defined SQL queries to select the data, but Genie struggles to use that data to actually plot the charts.
 
Is it better to let Genie calculate everything “on the fly” from raw tables? Or should I create a notebook with reusable SPC logic and have Genie call it when needed?
If the notebook approach is best, how can I make it easy for future users to just ask Genie in plain language (e.g., “Generate an SPC chart for measurement x”) and have it automatically use that notebook?
 
Thanks in advance for the help!
0 Kudos

curious_rabbit
New Contributor II

‎12-12-2025 12:05 AM

Or here is hopefully a more elegant way to phrase my question:

To visualise a control diagram in Genie for an end-user, should I a) instruct Genie how to create an SPC chart with SQL on the fly, of b) create a background job (pre-defined SQL query in the instructions, or a notebook) that prepares the data for charting and instruct genie to use that data?

0 Kudos

AbhaySingh
Databricks Employee
Databricks Employee
In response to curious_rabbit

‎12-23-2025 05:59 AM

Pre-computed control chart data (Option B) is strongly recommended over on-the-fly SQL generation.

While Genie can theoretically generate complex statistical SQL, the reliability ( or greater accuracy)  required for quality control systems demands pre-computed control limits, zone classifications, and violation flags.

0 Kudos