Data modeling in Analytics
Data modeling in Analytics is mapping the existing tables in your data warehouse into Analytics datasets. Analytics' warehouse native architecture makes it easy for you to leverage and reuse the existing schemas, tables, and governance for your business's product analytics needs. There is no need to configure new pipelines, copy data to third-party storage, or migrate existing data to rigid schemas.
This guide helps you with the following:
- Creating new connections to warehouses.
- Configuring datasets in Analytics according to best practices.
- Ensuring source tables or views meet warehouse requirements.
Rich text support for column, event, and dataset descriptions
Data warehouses support adding free-text comments or descriptions to columns and tables. This feature is available for all data warehouses – Snowflake, BigQuery, Amazon Redshift, and Databricks. Analytics automatically inherit the following description fields from the data warehouse:
Column descriptions
- Snowflake, Amazon Redshift, and Databricks use the COMMENT field for each mapped column in the data warehouses to automatically pull semantic column descriptions, and BigQuery uses the DESCRIPTION field.
- Source datasets inherit column descriptions directly from the corresponding column descriptions in the warehouse.
- Union datasets inherit descriptions from a column with the same name in one of the child tables. Typically, columns with the same name across child tables share the same description. No column descriptions are inherited for other dataset types because there is no direct correlation between warehouse columns and Analytics columns.
Table descriptions
- For source datasets, descriptions are inherited directly from the table descriptions in the warehouse.
- For other dataset types, no descriptions are inherited because there is no direct correlation between warehouse tables and Analytics datasets.
Event descriptions
- For union event datasets, if a specific child table represents an event type, the warehouse description for that child table is used for the corresponding event type.
- For other dataset types, no event descriptions are inherited because there is no direct correlation between warehouse and Analytics events.
The text is refreshed nightly from the data warehouse for all description types mentioned. However, if a user manually updates a description, the automatic refresh is disabled. If the user deletes the manual description, warehouse descriptions are reapplied.
Before you get started
- Ensure you are provisioned with an Analytics account and can log in. If you need help, reach out to your Analytics administrator.
- Go to the correct data app or create a new one.