Dataset Basics
Outline
Welcome
In Optimizely Analytics, Datasets are the foundation for all types of analysis—whether you're tracking events, segmenting users, measuring retention, or exploring data trends. Datasets allow you to bring in structured data from your data warehouse and define how that data should be interpreted inside Analytics.
This chapter will guide you through the different types of datasets available, how they’re created and managed, and how they support flexible, scalable, and real-time analytics use cases. You'll also learn how datasets map to warehouse tables, how to enrich them with metadata, and how to create powerful derived and union datasets to fit your product and business needs.
After completing this lesson, you should be able to:
Define what a dataset is in Optimizely Analytics
Understand the four main dataset types and their use cases
Create a dataset from a connected data warehouse
Choose the appropriate dataset type for your reporting needs
Configure columns, filters, and schema settings during dataset creation
What Is a Dataset?
A dataset is a structured selection of data from your warehouse that serves as a reusable source for reports, dashboards, and custom visualizations in Analytics. Datasets allow you to focus on relevant data while maintaining performance and governance.
Actor and Event Data
In Optimizely Analytics, datasets are logical views of your warehouse tables or views. They are used to model two key types of data: event data and actor data.
Event Data
Event data captures actions that occur at a specific point in time—like a page_view or purchase. These events can come from internal systems, third-party platforms, marketing tools, or customer care sources. Each event is linked to an actor, such as a user, account, or vendor who triggered the action.
Actor Data
Actor datasets represent entities (usually users) that perform events. They are typically mapped one-to-one from existing tables or views in your warehouse. Each row contains a unique identifier (like a UUID) and attributes that describe the actor. No special rules are needed to define actor datasets.
Dataset Types in Analytics
Analytics supports four distinct types of datasets. Each type has its own behavior, configuration options, and ideal usecases.
There are four types of datasets in Optimizely Analytics:
Source dataset
Derived dataset
Union dataset
Column actor dataset
Edit, duplicate, or delete a dataset
Check out the demo to understand how to edit, duplicate, or delete a dataset: https://optimizely.navattic.com/uzh807a1