What does the Data Product Category values of Foundational/Derived really mean??

Summary: Brian Lloyd initiated a discussion about the Data Product Workflows from the Collibra Marketplace, specifically questioning the meaning of 'Foundational' and 'Derived' categories for data products. victormorisset responded by explaining that 'Foundational' data products involve data close to the original source, whereas 'Derived' data products are built on top of other data products, often for analytical purposes. Brian Lloyd appreciated the clarification, noting that the message from victormorisset was far clearer than the original Collibra documentation.

Hello fellow Data Product aficionados!
We are playing around with the Data Product Workflows from the Collibra Marketplace. The requesting of a new Data Product has a question to populate the Data Product Category attribute with one of the two values of Foundational/Derived.

Looking at the Collibra documentation it is not clear what these two values mean (fundamentally) and how are they to be used? I would love to hear from Collibra Product Manager on what the attribute is for and what their fuller definition is for these two values.

Thanks!

-Brian

Brian Lloyd

Posted 1 year ago·Last reply 1 year ago

2 comments

Victor Morisset

·1 year ago

@Brian Lloyd

You can find the following description directly in the Collibra instance on the attribute type:

This attribute type helps categorizing the data products in for example more business user specific derived data products, versus more technical/foundational base data products targeted at the more technical users that are building the derived data products.

This attribute is meant to identify and differentiate between 2 types of Data Products:

Foundational: Data Products exposing data close to the original source. This is typically the information you would found in the landing zone or maybe the curated zone of a data lake.
Derived: Data Products that are built on top of other Data Products. It could be an analytical Data Products for example, that is aggregating the information across multiple dimensions to make it consumable for analytic purposes.

The overall idea is to be able to understand if this Data Product represent the most detail level of information we can get or if it was aggregated, filtered or transformed for a specific purpose, that conducted to a loss of information in the way.

It's important to understand depending on what you are looking to achieve.

Brian Lloyd

OP1 year ago

Thank you! I read the description attached to the attribute type om Collibra but defining the values buy using the values in the definition is a poor way to go about it.
What you added in the message makes FAR more sense. Thanks!!