3.1.1 Customer and Account Modeling Sandbox Workspace

The Cust Acct Sandbox is a pre-configured sandbox workspace for creating Customer and Account models. It includes a batch that populates data from the Production workspace into the Sandbox workspace. Users can utilize these datasets in pre-seeded templates to build models or scenarios.

The Cust Acct Sandbox workspace is configured with a pre-seeded batch that populates data into the FCC_ACCT_SMRY_MONTH and FCC_CUST_SMRY_MONTH tables from the schema of the Production workspace.

Manage Batch

To manage batches in the Cust Acct sandbox workspace, follow these steps:
  1. On the Workspace Summary, click Cust Acct Sandbox Workspace.
  2. From the Orchestration menu, select Schedule Batch. The Schedule Batch page is displayed.
  3. From the drop-down list, select the CustAcctSandboxPopulation batch.
  4. Click Edit Parameters to modify the dynamic parameters. The Edit Dynamic Params pane is displayed.
  5. Click CustAcctSandboxPopulation batch to view the Batch Parameters.
  6. Set the $BatchDate$ to define the batch execution date as follows:
    • Set the batch date to SYSDATE (system date). By default, the batch execution date is SYSDATE.
    • Toggle and select MISDATE from the Select Date picker to specify a particular batch execution date.
  7. Click Execute to run the batch.
    You can click Preview Jobs at any time to view the process sequence used to execute the selected batch.

    Note:

    For more information about Schedule Batch, see the Managing Batch/Batch Group Executions section.
  8. From the Orchestration menu, select Monitor Batch. The Monitor page is displayed.
  9. Click Start Monitor to track and monitor batch process executions.

    Note:

    For more information about the Monitor Batch, see the Monitor Batch/Batch Group section.

Accessing the Supervised Training Notebook

The Cust Acct Sandbox workspace contains the Supervised Training notebook; a sample notebook that demonstrates the full workflow required to build a Supervised Machine Learning (ML) model. The Notebook starts with the selection of the dataset, performs necessary preprocessing and transformation on the raw data and trains an ML model based on this data.

This trained model is persisted into the database for the scoring in Production to refer. During scoring, the model is retrieved and SAR scores are predicted for each entity in the production data and the scores are persisted in an Output table.

To access the Supervised Training Notebook, follow these steps:
  1. On the Workspace Summary, click Cust Acct Sandbox Workspace.
  2. From the Modeling menu, select Pipelines. The Model Pipelines page is displayed.
  3. Click the Supervised Learning folder, then click Supervised Learning Example objective. The Pipeline tab is displayed.
  4. Click the Notebook tab.
  5. At the top of the notebook, click Run Paragraphs to execute all paragraphs.
  6. Navigate to the Read and Sample Data paragraph and view the account and customer summary month data, which can be used to build account-focused or customer-focused models for Transaction Monitoring.