Data Integration and Analytics
Domain 5 asks you to select the right service for a data movement, query, or model lifecycle requirement.
Service fit
- Data Virtualization: query remote sources without copying; good for federation and rapid access.
- DataStage: design ETL/ELT pipelines, transformations, and scheduled movement.
- Db2 Warehouse: warehouse workloads and SQL analytics.
- Watson Studio: notebooks, visual modeling, and project collaboration.
- Watson Machine Learning: deploy and monitor models as online or batch deployments.
- Analytics Engine powered by Apache Spark: large-scale Spark jobs.
- Watson OpenScale: model quality, fairness, and drift monitoring.
- Match 360: entity resolution and master data.
Worked example
A team needs a single customer view across CRM, ERP, and a lakehouse, with matching across name and address variants. Data Virtualization gives a query layer but does not resolve identities. Match 360 is the fit for entity resolution; DataStage or Data Virtualization can feed it. State the requirement first, then map the service.
Traps
Using Watson Studio as a production BI server. Assuming Data Virtualization copies data into Cloud Pak for Data. Deploying a model in WML without a deployment space and monitoring plan. Sizing Spark jobs from notebook performance.