Data Modeling

Welcome​

​In this lesson, you'll learn how data modeling works in Optimizely Analytics and how it leverages your existing warehouse setup. With its warehouse-native architecture, Analytics lets you define datasets directly from your current schemas—without moving data, building new pipelines, or restructuring your models.​

​After completing this lesson, you should be able to: ​

• Understand the concept of data modeling in Analytics ​

• Create new connections to data warehouses ​

• Configure datasets using best practices ​

• Identify how descriptions for columns, tables, and events are inherited ​

• Understand how Analytics synchronizes metadata from your warehouse

What is Data Modeling in Analytics?​

​Data modeling in Analytics refers to the process of mapping existing tables in your data warehouse into Analytics datasets. Since Analytics is warehouse-native, it allows you to reuse your existing schemas, tables, and governance—eliminating the need for creating new data pipelines, copying data into external systems, or transforming it to fit rigid schemas.

Setting Up Your Data Model​

​To begin data modeling in Analytics, you'll need to:​

1. Create New Connections to Warehouses​Establish a connection between Analytics and your cloud data warehouse to start mapping tables.​

2. Configure Datasets According to Best Practices​Set up datasets in a way that aligns with your organization’s data usage patterns and Analytics’ recommended configuration guidelines.​

3. Ensure Source Tables or Views Meet Requirements​Make sure that the tables or views from your warehouse are structured in a way that meets Analytics compatibility standards.​

Description Inheritance in Analytics​

Analytics supports rich text descriptions for columns, events, and datasets. These descriptions are inherited directly from the data warehouse—enabling clarity and consistency across platforms.​

​Column Descriptions​

Column descriptions behave differently depending on the dataset type:​

• Source datasets automatically pull column descriptions from their corresponding columns in the data warehouse.​

• Union datasets inherit descriptions from columns with the same name found in one of their child tables.​

• Other dataset types do not inherit column descriptions, as there is no one-to-one mapping with warehouse columns.​​

Table Descriptions​

Table descriptions follow similar rules:​

• In source datasets, table-level descriptions are inherited directly from the related tables in the warehouse.​

• Other dataset types do not support inherited table descriptions due to the absence of a direct warehouse mapping.​​

Event Descriptions​

Event descriptions are supported only under specific conditions:​

• Union event datasets will inherit event descriptions when a child table clearly represents a distinct event type—the description from that warehouse table is applied.​

• Other dataset types cannot inherit event descriptions, as no direct relationship to warehouse events exists.

​Note: All inherited descriptions are refreshed nightly from the warehouse. However, if a user manually edits a description in Analytics, the automatic refresh for that field is turned off. Deleting the manual description will re-enable automatic inheritance.​

Before you start make sure of the following:​

• You have an active Analytics account and can log in successfully​

• You are in the correct data app or have created a new one for your datasets​

​If you run into access issues, contact your Analytics administrator.​

To explore more about how datasets are structured and mapped within Optimizely Analytics, you can refer to the Data Modeling in Analytics documentation for a detailed guide on best practices, setup steps, and supported warehouse configurations.​