Edit

Ontology (preview) tutorial part 3: View the ontology

In this tutorial step, explore your ontology by viewing the entity type details included in ontology (preview). Inspect entity instances that instantiate your entity types with data, and explore relationship graphs that provide context across sales and device streaming data.

Important

This feature is in preview.

View entity instances

When you bound data to your entity types in previous tutorial steps, ontology automatically created instances of those entities that are tied to the source data rows. In this section, you observe those entity instances and their data.

View instance list and static data

  1. Start in the Home configuration canvas of ontology. Select the SaleEvent entity type, and View Entity Type details from the top ribbon.

    Screenshot of opening the entity type details for SaleEvent.

  2. Open the Instances tab. Verify that it shows six entity instances with data populated from the factsales lakehouse table, like revenue and unit counts.

    Screenshot of the Freezer instances.

    Tip

    If data bindings don't load, confirm that the source data tables exist with matching column names, and that your Fabric identity has data access.

View time series data

  1. In the top left corner of the page, use the selector next to the entity type name to switch to the Freezer entity type.

    Screenshot of switching to a different entity type.

  2. Open the Overview tab. The tab loads with empty charts, because the default time range of Last 30 days doesn't include any data.

    Screenshot of the Overview tab.

  3. Update the time range from the default of Last 30 days to a custom date range that begins on Fri Aug 01 2025 at 12:00 AM, ends on Mon Aug 04 2025 at 12:00 AM, and has a Time granularity of 5 minutes.

    Screenshot of the time selector.

  4. Observe the time series data that's now visible from several Freezer entity instances in the time window you selected.

    Screenshot of the time series tiles.

View ontology graph

The Overview tab also contains a Relationship graph, which you use to visualize your ontology in a graph of nodes and edges.

  1. Use the entity type selector to switch to the SaleEvent entity type. In the Relationship graph tile, select Expand.

    Screenshot of expanding the SaleEvent relationship graph.

  2. The expanded graph view opens. Observe the details of the relationships from the SaleEvent entity type to Products and Store.

    Screenshot of data in the expanded relationship graph.

  3. Use the entity type selector to switch to the Store entity type. Expand its relationship graph.

    Screenshot of expanding the Store relationship graph.

  4. In the graph, observe the relationships that Store has with Freezer and SaleEvent. Then, select Run query in the query builder ribbon. This action runs the default query and shows a graph of entity instances alongside their connections.

    Screenshot of the Store relationship graph and instances.

    Tip

    If the graph result looks sparse, check the entity type keys in the data bindings and verify that they match the keys defined in Create entity types and data bindings. For example, the key for the SaleEvent entity type is SaleId.

Query graph instances

In the relationship graph view, you can query your ontology for entity instances that meet certain criteria. Use the Query builder filters in the top ribbon to craft queries.

Screenshot of selecting the query builder.

First, craft this query: Show all freezers that are operated in the Paris store.

  1. In the Store entity's relationship graph, select Add filter > Store > StoreId from the query builder ribbon. Set the filter for StoreId = S-PAR-01. This value is the store ID for the Paris store.

    Screenshot of filtering by Store ID.

  2. In the Components section, uncheck SaleEvent so that the only checked fields are Nodes > Store, Nodes > Freezer, and Edges > operates.

    Screenshot of filtering the components.

  3. Select Run query and verify that the instance graph shows two freezers connected to the Paris store.

    Screenshot of the freezers that are connected to the filtered store.

  4. Select Clear query to clear the query results.

Next, craft this query: Show all stores that have made a sale with a revenue greater than 150.

  1. Select Add a node and add a node for SaleEvent.

    Screenshot of adding nodes to a new query.

  2. In the Components section, check the boxes next to Nodes > Store and Edges > from to add them to the graph.

  3. From the query builder ribbon, select Add filter > SaleEvent > RevenueUSD. Set the filter for RevenueUSD > 150.

    Screenshot of filtering by sale revenue.

  4. Select Run query and verify that the instance graph shows two stores that meet the filter for their connected sale events. You can also select the nodes in the graph to get details of the specific sale events.

    Screenshot of the stores that meet the filter for their connected sale events.

This process allows you to inspect the paths that connect operational issues (like rising freezer temperature at certain stores) to business outcomes (sales).

Next steps

In this step, you viewed the instances connected to your ontology and explored the data they contain. Next, create a data agent to explore the data further by using natural language queries.

Continue to Create data agent.

  • Last updated on