Skip to main content

Modeling, Products & BI

This area follows a dataset from the moment it is shaped to the moment it is consumed. You model it, you package it as a product with a contract, the catalog records what it is and where it came from, a reporting tool reads it over a protocol it already speaks, and (when you choose) it becomes a listing on a cross-organization marketplace. Every stage runs on the same substrate, so a read still carries its lineage, its classification, and its cost.

In this area​

PageWhat it covers
Data ModelingThe transformation-project modeling layer: projects, models, and seeds compiled into a governed DAG
The semantic layerMetric definitions authored once and emitted as native configuration for the BI tools your analysts already run
Data productsWhat a data product is: typed ports, a descriptor, certification, provenance, and the path to the exchange
Reporting & BI tool supportEvery surface a reporting tool connects over: Postgres wire, Delta Sharing, Iceberg REST, ODBC/JDBC, GraphQL: governed the same way

The through-line​

  • A model is governed the moment it runs. Data Modeling compiles into a DAG whose every read passes the same in-engine enforcement as any other query. There is no unpoliced build path.
  • A product is more than a location. A data product carries typed ports and a contract, so a consumer knows the shape and the guarantees, not just where the bytes live.
  • The catalog answers from the cell. It is replicated, distributed, and federated, so "what data do we have and where did it come from" is a local answer, not a remote lookup. See the catalog and lineage.
  • Existing tools connect unchanged. A reporting tool reaches the platform over a standard protocol, and the policy is already in the query plan the same policy whether the caller is a BI tool or an agent.