Testing the Prediction Studio APIs

Pega provides APIs that allow you to connect to Prediction Studio from your preferred data science environment. These APIs provide endpoints for many Prediction Studio operations.

To provide data scientists with the option to work from their preferred Python environment, a System Administrator generates the artifacts required to export ADM snapshots and creates the necessary credentials to access the APIs.

To generate the data sets and data flows required for the export of ADM data, you switch to the settings of Prediction Studio and configure the Monitoring database export. Select the Ruleset for the export artifacts and submit to complete the configuration.

Back in Infinity Studio, create new client credentials for access to the Prediction Studio APIs. You create an OAuth 2.0 client registration data instance. After creating the client registration, ensure to save the credentials that include the Client ID and Client Secret.

To inspect the API responses, you open the API landing page. From here, select the Prediction APIs service package and authorize the client by entering the saved Client ID and Client Secret.

The tasks are divided into categories like Data Export, Models, Reporting, Predictions, and Prediction Studio. Explore the available models by using the GET List models API to retrieve a list of models within the application.

To delve deeper into a specific model, such as the Web Click Through Rate Customer adaptive model, use the GET Describe Model API with the appropriate model ID to obtain detailed information, including the modeling technique, target labels, and predictors. This model calculates the propensity that a customer clicks on a web banner.

To validate the generated artifact required for the export of ADM snapshots, you use the POST Trigger Datamart Export API. In the repository, the model snapshots and predictor snapshots are stored in the datamart folder. The APIs are now ready for use by the data science team.