Developer Portal
Documentation
Tools
Expert Center
Introduction to Expert Center
HierarchyUnderstanding EntitiesSeven Stages to EnrichmentGetting Started
Step-by-Step Building Onboarding Guide
Step 1: Pull
Step 2: Label
Step 3: Process
Step 4: Review
Step 5: Unify
Step 6: Push
Step 7: Validate
FAQs - Expert Center
Connector Guides
Gateway Setup & Administration
Support & Policies
API ReferenceConsole

Introduction to Expert Center

Expert Center combines AI-powered automation with targeted human expertise to create knowledge graphs for the built world in days instead of months.

Hierarchy

Expert Center organizes data in a three-level hierarchy:

Embedded hierarchy visualization showing Expert Center as a rectangle containing an Organization, Building, and Connector.

Organization → Building → Connector

  • Organization: Your company or client account
  • Building: Individual facilities within the organization
  • Connector: Data sources within each building (BMS, IoT, Engineering Drawings, etc.)

Understanding Entities

An Entity in Mapped is a distinct object, concept, or space within a building that serves as a core unit for organizing and representing data. Entities can represent anything from sensor readings and equipment to rooms and logical groupings, and they include attributes like name, description, object type, and more. Most commonly, entities are Points, Things, Places or Collections.

We'll review different types of entities in more detail for each of the seven stages to enrichment. Read more about Entities in the FAQs.

Seven Stages to Enrichment

Diagram of enrichment workflow. The knowledge graph is represented as a teal cylinder above the Expert Center hierarchy. The first step is an arrow exiting the Knowledge graph labeled Pull, which connects to a box labeled Label the second step inside the Connectors box. Then the arrow continues to a box labeled Process the third step, then a box labeled Review the fourth step, then a box labeled Unify the fifth step. Finally an arrow exits Unify and returns from the other side to the Knowledge graph for step six Push. From the top of the Knowledge Graph, step 7 exits as an arrow that diverges to Independent Solution providers and app.mapped.com

1. Pull

When you Pull, you ingest an organization’s data from the knowledge graph. Continue reading through the Pull Step.

  • Input: An organization’s knowledge graph with raw data
  • Output: Buildings and connectors loaded into Expert Center with their associated entities

2. Label

Through labeling, you'll manually enrich entities with classifications, derived entities, and relationships between entities. Continue reading through the Label Step.

  • Input: Raw entities (e.g., "VAV1_ZNT")
  • Output: Labeled entities with:
    • Classifications (e.g., Zone Air Temperature Sensor)
    • Derived entities (e.g., VAV1)
    • Derived Relationships (e.g., VAV1 has a ZNT point)

Diagram of relationships between VAV1 which is of Variable Air Volume Box type, and has a Point VAV1.ZNT which is of Zone Air Temp Sensor type.

3. Process

With this step, you'll apply AI models to process remaining entities. Continue reading through the Process Step.

  • Input: ~0.5-5% labeled entities are used as training data
  • Output: AI predictions for all entities include confidence scores (Low/Medium/High)

4. Review

Through Review, you'll evaluate AI predictions and correct any errors. Continue reading through the Review Step.

  • Input: Inferred entities with confidence scores
  • Output:
    • Approved inferences (confirmed correct)
    • Declined inferences (sent back to labeling queue)

Based on the review results, you may need to return to the labeling stage to improve model accuracy.

5. Unify

Unification merges duplicate entities across data sources. Continue reading through the Unify Step.

  • Input: Inferred Source and Derived entities from a building’s connector(s)
  • Output: Single unified entities representing physical equipment. For example, an entity known as AHU1 from your BMS combined with "Air Handler 1" from the CSV source connector becomes one unified AHU1.

Diagram of the unification process where metadata about VAV1 and VAV101 from Connectors A and B becomes one Unified Entity.

6. Push

The Push step sends enriched data to the knowledge graph. Continue reading through Push Unify Step.

  • Input: Model in Expert Center
  • Output: Knowledge Graph updated with changes made in Expert Center

7. Validate

You'll complete this process by inspecting enriched data in the knowledge graph.

  • Input: Updated knowledge graph with enriched building data
  • Output: GraphQL query results confirming data structure and relationships

Getting Started

Ready to begin?

Follow our Step-by-Step Building Onboarding Guide for detailed instructions on implementing each phase within Expert Center, or continue with Step 1: Pull.