Our Ontology Builder: Modeling Your Business

Our Ontology Builder is where the most critical step of your analytics journey begins. Before you can build a single trusted Workbook or ML model in ML Hub, you must first create a clean, reliable map of your business. This is that tool.

Here, you will transform raw, physical data tables into a meaningful "digital twin" of your enterprise. You will create reusable Objects (e.g., Customer, Product) and define their Relationships that become the single source of truth for the entire Arkham platform.

How It Works: A Tour of our UI

Our Ontology Builder is structured around a central visual graph and three key panels for inspection, management, and impact analysis.

1. Our Ontology Graph

The central part of the screen is our Ontology Graph, a canvas that visually represents your Objects and the relationships between them.

Object Nodes: Each box in the graph is an Object (e.g., Pilot, Aircraft, Flight). These are the core entities of your business.
Relationship Links: The dotted lines connecting the nodes represent the Relationships between objects. The forks in the lines indicate the cardinality of the relationship (e.g., one Aircraft can have many Flights).

You can click and drag to pan the canvas and use the mouse wheel to zoom in and out, making it easy to navigate even very large and complex ontologies.

2. Details Panel

When you select an Object in our Ontology Graph (like the Aircraft Object in the example), our Details Panel appears at the bottom. This is where you can inspect and manage the selected Object.

Our Preview tab allows you to see the live data for the first 100 instances of the selected Object, directly from the Lakehouse. This is invaluable for quickly validating that your Object is mapped correctly and that the underlying data is as you expect, before it gets used in a critical report.

Our Details tab provides a comprehensive overview of the Object's configuration.

Description: A human-readable description of what the Object represents. This context is surfaced across our platform to help users understand the data they are working with.
Properties: A list of all the attributes that define the Object. Each property has a business-friendly name (e.g., Year, Built) and a description. These become the trusted, reusable dimensions and features for downstream analytics and ML models.
Relations: A list of the formal relationships this Object has with other Objects. This is the core of the semantic model, enabling you to traverse your business graph and ask complex questions.

3. Dependents Panel

On the right-hand side, the Dependents Panel provides critical, automated data lineage. It shows you all the downstream resources in our Arkham platform that consume the selected Object. This is your primary tool for impact analysis.

Workbooks: Shows a count of all Workbook that visualize this Object. This prevents you from breaking dashboards that rely on this part of the business model.

This panel removes the guesswork from maintenance, making your semantic layer safe to evolve and improve over time.

4. Key Actions

Add Object: The + Add Object button in the top-right corner opens a wizard that allows you to create a new Object by mapping it to a production-grade dataset in your Lakehouse.
Edit / Add Relation: Within our Details Panel, these buttons allow you to modify the selected Object's metadata or define new relationships with other Objects on the canvas, enriching the semantic model.

Key Technical Benefits

Visual Business Modeling: Our Ontology Graph in a canvas UI makes our semantic layer intuitive and easy for all stakeholders to understand.
Automated Impact Analysis: Our Dependents Panel provides instant, automated lineage, showing you exactly what will be affected by a change before you make it.
Live Data Previews: Instantly validate your object mappings against live data from our Lakehouse without leaving our UI.
Centralized Business Logic: By defining objects and relationships here, you create a single source of truth for business concepts that is reused across our entire Data Platform and AI Platform.

🤖 AI-Assisted Ontology Modeling with TARS

As your AI co-pilot, TARS can significantly accelerate the development of your Ontology. Instead of manually mapping objects and properties, you can use natural language to have TARS perform these actions for you. It can also help you understand the current state of your ontology and its downstream dependencies.

"TARS, analyze the @sales_transactions and @customer_details datasets. What objects and relationships could I model from them?"

Ontology Overview: Understand how Objects fit into our broader semantic layer.
Data Catalog: The source of the trusted Production Datasets used to create your Objects.
Workbooks: See where your Objects and their properties are consumed in downstream dashboards.