Data modeling in Data Visualization

With Cloudera Data Visualization, you can create logical datasets that model semantic relationships across multiple data sources. These datasets enable you to build blended visuals, dashboards, and applications.

You can expand a basic, one-table dataset by creating joins with other relevant tables from the same source or from other data stores. The joins created in Data Visualization are not materialized. They are calculated during run-time.

Combining data across multiple tables enriches the dataset considerably; it enables you to conduct much more meaningful research and produce insightful visual analytics.

There are two distinct approaches to create data joins for visualization:

• Defining in UI is ideal for datasets that include star-type schemas.
• Defining on back-end ETL is preferable for fact-fact joins, so they may be pre-materialized.

See Working with data models for instructions on how to create and manage joins.

See topics in How To: Advanced Analytics for more advanced data modeling details, such as dimension hierarchies, segments, and events.

See Setting filter associations for information about filter associations related to data modeling and management.