Einstein Prediction Services using Postman: Einstein Discovery

Utilize the Einstein Prediction Service to obtain predictions through your preferred REST client.

Step A: Establish a Connected Application within the Salesforce platform.

• Navigate to the Quick Find Box and go to “Apps” and then “App Manager.”
• Create a new Connected App.
• Specify a name for the Connected App under “Connected App Name.”
• Input the email address into the designated field labeled “Contact Email.”
• Check the “Enable OAuth Settings” checkbox.
• Set the Callback URL to https://login.salesforce.com/
• Navigate to the “Selected OAuth Scopes” section and select both “Access Analytics REST API resources (wave_api)” and “Manage user data via APIs (api).”
• Save your settings and then click “Continue.”
  
  

• Navigate to “Manage Consumer Details” and retrieve both the Consumer key and secret.
• Access the “Connected App” section, proceed to “Manager,” and then select “OAuth Policies.”
  
  

Step B: Configure authentication in the REST Client to obtain the Access Token and Instance URL

Setting up authentication in your REST client using the recently created Connected App credentials is vital for accessing the Einstein Prediction Service. Authentication is a mandatory step for utilizing the service.

How to Generate Access Token and Instance URLs?

• Open your preferred web browser and log into Salesforce. In a new tab, enter the following request URL:
• https://<YOUR_INSTANCE>.salesforce.com/services/oauth2/authorize?response_type=code&client_id=<CONSUMER_KEY>&redirect_uri=https://login.salesforce.com/
• If prompted, allow access. The browser will redirect to a URL similar to the following, from which you can obtain the authorization code:
• Open Postman.
  
  

• Retrieve the access token and instance URL using the obtained auth.
  
  

Step C: Retrieve a list of Prediction Definitions currently available.

• Create a new GET request and name it as “Get Prediction Definitions.”
• Set the Request Type to GET and the Request URL to instance_URL/services/data/v55.0/smartdatadiscovery/predictionDefinitions.
• Navigate to the Authorization tab.
• Choose Authorization Type as OAuth 2.0.
• Insert the access token into the designated field labeled “Set Access Token” in your document.
• Save the request configuration.
• Click the “Send” button to execute the request.
  
  

Step D: Obtain metadata for a specific Prediction Definition.

• In Step 3, the response contains the URL for fetching metadata related to the prediction definition created earlier. Please create a new GET request, name it “GET Prediction Definition,” and replace the URL with the following string:
• https://InstanceUrl/services/data/v55.0/smartdatadiscovery/predictionDefinitions/ID
• Note that ID is the identifier copied from the previous “Set Up Predictions in Salesforce” step.
• Execute the request by clicking “Send” and examine the resulting response.
  
  

Step E: Retrieve a list of Models associated with a particular Prediction.

• In Step 4’s response, there is a URL provided for obtaining metadata related to the previously deployed model. Introduce a new GET request named “Get Models” and replace the existing request string in the URL with the following:
• https://InstanceURL/services/data/v55.0/smartdatadiscovery/predictionDefinitions/ID/models
• Execute the request by clicking “Send” and analyze the resulting response.
  
  

Step F: Acquire predictions using the specified methodology.

• Create a new request named “Get Predictions” with the following configurations:
• Request Type: POST
• Request URL: instance_URL/services/data/v55.0/smartdatadiscovery/predict
• Configure the Authorization Settings with the appropriate Authorization Type and Access Token.
• For the POST call, define the request body as follows:
• Body Type: RAW
• Body Type Request: JSON
• Include the following JSON-formatted request body:
  
  

{
“predictionDefinition”: “1OR5j000000TPQ5GAO”,
“type”: “RawData”,
“columnNames”: [“Quantity”, “Category”, “Sub_Category”, “Sales”, “Profit_per_Order”],
“rows”: [
[“2”, “Furniture”, “Chairs”, “262”, “42”]
],
“settings”: {
“maxMiddleValues”: 2,
“maxPrescriptions”: 2
}
}

• Ensure that you replace “yourPredictionDefinitionId” with the actual Prediction ID obtained earlier (ID).
• The “maxMiddleValues” parameter is set to 2 to enable the return of top factors, and “maxPrescriptions” is set to 2 to enable the return of improvement suggestions.
• Save the request, click “Send,” and review the response.
  
  

In this context:

“Total” signifies the predicted value.

The “Middlevalues” segment outlines the primary influencing factors associated with this prediction.

Illustration:

• When the Sub Category is “chairs” and the Sales amount is 262, it results in a deduction of approximately 731 from the prediction (-730.8852).
• If the Sales amount is 262, it leads to an elevation in the prediction by approximately 245 (244.8200).