Aggregate Dimension
A special type of feature dimension transformation is aggregated before its joined to the activity stream.
What is an aggregate dimension table?
An aggregate dimension table is a special type of feature dimension table that is aggregated before its joined to the activity stream using feature values specified by the user during dataset assembly. Any metrics stored in this table are aggregated to the level of detail of the selected feature before the join is performed.
How is it used?
It is commonly used to track marketing spend because impressions, clicks, and spend cannot be tied to a customer or customer action. Instead these metrics are associated with campaign details (utm_source, utm_campaign, ad_set, etc).
See How To: Add Spend Data to your Dataset to see how it's used.
Columns in an Spend Transformation
A spend transformation generates a dataset with six required columns and any number of additional columns related to marketing (campaign, utm parameters, etc). Spend transformations are a specific implementation of an enrichment transformation.
| Column | Description |
|---|---|
id | Unique id for each record in the table |
ts | Timestamp in UTC of the data date |
| additional__dimension_columns | These columns are the additional features that can be used as join keys when added to the dataset. The names and formatting should match those in the activities. The names should be descriptive of the data they represent and do not need to use the feature_ naming convention that activities use. |
| additional__metric_columns | These columns are numerical columns that are aggregated via SUM() before they are joined to the activity stream. The names should be descriptive of the data they represent and do not need to use the feature_ naming convention that activities use. |
How To: Create a Spend Table
Watch this step-by-step tutorial to learn how to add an aggregate dimension table.
Examples
Explore the transformation library for examples by data source.
FAQ
Still have questions?
Our data team is here to help! Here are a couple ways to get in touch...
💬 Chat us from within Narrator
💌 Email us at [email protected]
🗓 Or schedule a 15 minute meeting with our data team
Updated 11 months ago