Dimensions

A Dimension is a column that experiment results can be sliced by in the Explore analysis. Dimensions live under Data Management → Dimension.

Where to find this in the console

Data Management → Dimension. The page has a search box (Search by name) and a Create button.

Columns shown

Column	What it shows
Name	The dimension name, tagged with its type (e.g. User).
Data Source	The table and column it reads, e.g. abc_user_property_date_partition.country.
Dimension Groups	Buckets defined on the dimension, e.g. non_payer;minnow;dolphin;whale.
Owner / Created at	Owner and creation time.

The only row action today is Delete.

Create a dimension

Create opens the Create explore dimension dialog:

• Name — the dimension name.
• Data Source + Fields — the source table and the column to read.
• Dimension Type — "Determined by the data source", one of:
  • User Attribute — for example, name, version number, app type (read from a User Property Table).
  • Event Type — for example, page ID, scene ID (read from a Fact Table).
• Dimension Groups (Optional) — use + Add Dimension Group to bundle several values into one bucket so Explore compares groups instead of every raw value.

Confirm with Confirm or discard with Cancel.

Where dimensions show up

• In the Explore view of an experiment result, where you pick which cut to compare.
• As candidate sub-populations in HTE (Heterogeneous Treatment Effect) Analysis.