Datarobot Score

Retrieves predictions from a DataRobot model that was deployed in Snowflake.

Parameters

Name	Type	Description	Is Optional
function_name	value	name of the custom Snowflake function that represents the DataRobot model
include_cols	column_list	List of columns to select
num_explains	value	(Optional) the number of prediction explanations to also retrieve. Prediction explanations are computationally expensive. If you supply num_explains, you must also supply threshold_low and threshold_high.	True
threshold_low	value	(Optional) predictions lower than this threshold will also calculate an explanation.	True
threshold_high	value	(Optional) predictions higher than this threshold will also calculate an explanation.	True

Example

toscore = rasgo.get.dataset(resource_key='adventureworks_sales_by_day_toscore')

# with explanations
scored = toscore.datarobot_score(function_name = 'PUBLIC.DEMO_PREDICT_NEXT_WEEK_SALES_V2',
                     include_cols = ['ORDERDATE','SALESAMOUNT'],
                     num_explains=2, 
                     threshold_low=100, 
                     threshold_high=100000)

# without explanations
scored = toscore.datarobot_score(function_name = 'PUBLIC.DEMO_PREDICT_NEXT_WEEK_SALES',
                     include_cols = ['ORDERDATE','SALESAMOUNT'])    

scored.preview()

Source Code

RasgoTransforms/datarobot_score.sql at main · rasgointelligence/RasgoTransformsGitHub