The Ontology
Module Summary
Understand the Ontology — Foundry's semantic layer that turns raw data into meaningful, interconnected business objects.
What Is the Ontology?
At its core, the Ontology is Foundry's semantic layer. It sits on top of your raw datasets and gives them meaning — turning rows in a table into real-world concepts like Employees, Orders, or Sensors.
Without the Ontology you have powerful data pipelines, but end users still need to know table names, join keys, and column encodings. With the Ontology, they search for "Employees in Berlin" and get exactly what they expect.
Why Does It Matter?
The Ontology is the single source of truth that every Foundry application reads from. Workshop apps, Functions, automated Actions, and even external APIs all reference the same Object Types and Link Types.
This means a change to the underlying pipeline automatically flows through to every dashboard, every alert, and every operational tool. There is no "sync" step — the Ontology is always live.
How Is It Built?
You construct the Ontology by:
1. Backing datasets — clean, transformed tables produced by your pipelines.
2. Object Type definitions — mapping columns to Properties (name, type, description).
3. Link Types — declaring relationships between Object Types (e.g., an Employee *works at* a Location).
4. Actions — defining safe, auditable operations users can perform (edit a property, create a new object, etc.).
Each layer builds on the one below, so well-modelled pipelines make for a clean Ontology.
Key Takeaways
- The Ontology turns raw data into meaningful, searchable business objects.
- Every Foundry application — Workshop, Functions, Actions — reads from the same Ontology.
- Object Types, Link Types, and Actions are the three core building blocks.
- A well-modelled pipeline is the foundation of a clean Ontology.