Perform contextualization

The contextualization feature within digital twin builder (preview) allows users to further augment the context of their data by creating semantic relationships between entities in their ontology.

Relationships are like the predicates in Resource Description Framework (RDF) triples, where each triple consists of a subject-predicate-object structure. For example, a triple might consist of a machine (subject) that is operated by (predicate) an operator (object). "Is operated by" is an example of a relationship. Relationships in digital twin builder describe how entities are linked to each other, and they play a fundamental role in providing context on data.

Here are some uses and benefits of creating semantic relationships in digital twin builder.

• Connecting entities. Relationships define how resources are connected, allowing you to create a network of data. For example, you can link a Person entity to an Organization entity with a worksFor relationship. This process allows your data to become a more complete representation of your real-world environment.
• Defining semantics. Relationships add semantic meaning to data, by defining the nature of connections between resources and processes in the real world. This process allows the ontology in digital twin builder to be more understandable and structured, which facilitates machine comprehension.
• Data interoperability and discoverability. Semantic relationships standardize connections across entities, making it easier for people to query and navigate data. It also helps downstream and consumer systems understand and integrate diverse data sources.

Prerequisites

• A workspace with a Microsoft Fabric-enabled capacity.
• Your desired data in a Fabric lakehouse with the necessary ETL done.
  • Time series data should be in columnar format. Columnar time series data is structured so that each column represents a different variable or attribute, while each row corresponds to a specific timestamp.
• A digital twin builder (preview) item with at least two entities mapped.

Create a relationship

This section describes how to add a new relationship between two entities in the semantic canvas.

1. In the semantic canvas, select one of the entities involved in your relationship, and select Add relationships in the menu ribbon. Or, select ... to the right of the entity name and then Add relationship from those options.

   

2. In the Relationship configuration pane that appears,

   1. Select the name of the First entity, and which Property to join in the relationship. Repeat the process for the Second entity.

   2. Define the relationship with a Relationship name.

   3. For Select relationship type, choose the appropriate cardinality.

      If a single source entity instance might be connected to many target entity instances, choose 1:N. If there are many source entity instances that might connect to a single target entity instance, choose N:1.

3. When you're finished adding the configuration details, select Create.

   

The relationship appears in the semantic canvas immediately. It might take a few minutes for it to propagate across the data layer and be available in the explorer.

Manage relationships

To manage an existing relationship, select it from the semantic canvas. You can change the selected entities and entity properties, and the relationship's cardinality.

To edit a relationship, follow these steps.

1. Find the relationship in the semantic canvas, by selecting its source or target entity from the Entities pane and selecting the relationship from the canvas.

   

2. Update the relationship configuration. When you're done with your changes, select Apply.

The relationship updates in the semantic canvas immediately. It might take a few minutes for it to propagate across the data layer and be available in the explorer.

Limitations

Relationships in digital twin builder (preview) have the following restrictions:

• All relationships must reference entity properties. This information is used by digital twin builder to create connections in the data layer based on specific reference keys (like in a traditional relationship database scenario).

• A single entity property can't be both the source and target in the same relationship.

  For example, say the Employee entity has the following properties: EmployeeId, EmployeeEmail, and ManagerId.

  The semantic relationship Employee (EmployeeId) -reportsTo-> Employee (ManagerId) is valid, but Employee (EmployeeId) -is-> Employee (EmployeeId) isn't valid.