Data products overview
Data products in Cortex Framework represent the curated, logical packaging of data assets engineered to solve specific business problems. They serve as the ultimate business-facing layer, directly powering business intelligence (BI) dashboards, reporting tools, machine learning models, and autonomous agents.
Dataform pipelines function as the execution engine for Cortex Framework available data products. They run extract, load, transform (ELT) processes that aggregate, join, and shape data from the standardized data foundation layer and, where applicable, other upstream data products into these final, consumer-ready structures.
Key concepts
Data asset - A pointer to a physical data resource, such as a BigQuery dataset, table, or view. A data product contains one or more assets.
Data product - A curated, logical grouping of data assets, formally packaged to be discoverable, trusted, and accessible for solving specific business problems. Cortex Framework delivers two types of data products:
Source-aligned data products: A collection of tables in an analytical data model (for example, sales orders or customers). They feed from the data foundation layer, serve as foundational entities, and can be used to feed more complex use case specific models.
Consumption data products: Specialized analytical models engineered for particular business scenarios, delivering complex KPI logic and high-fidelity metrics (for example, sales performance insights). Consumption data products typically consume and aggregate upstream data foundation modules, data products exposed by catalog modules, as well as source-aligned data products, to drive immediate decision-making.