Ontology Modeler Path

About this Path

This Path takes you through D.Hub portal from the ontology modeler point of view. If the Analyst Path is the flow of a person building result screens on top of data that already exists, and the Data Engineer Path is the flow of ingesting that data from external systems and automating it, this Path covers the third leg — modeling data as a graph of entities and relations — end to end across six lessons. About 55 minutes in total.

The persona is intentionally not narrowed. A data architect, a domain modeler, or a bridge between engineer and analyst can all enter, so the body is written in a task tone rather than a role tone — how to draw an entity, how to connect a relation, and how to inspect the instances that land on top.

Each lesson is 5–10 minutes. You can run through them in one sitting or spread one or two per day.

Terms to know before you start

• Ontology: A semantic model that defines business objects, their attributes, and connection rules.
• Entity: A uniquely identifiable object such as a machine or product whose attributes are recorded.
• Relation: A rule describing how two entities connect and in which direction.
• Identity Key: The attribute used to distinguish entity instances and update the same object consistently.
• Cypher: The graph query language used in Graph Explorer to retrieve entity and relation instances.

Prerequisites

• Portal access with the Editor role or above in the practice collection. You need to be able to create entities, relations, datasets, Code, and pipelines.
• Familiarity with collection and dataset basics. The depth from Finding data — Collections and search and Create your first collection and register a dataset is enough.
• An empty practice collection is sufficient. Lessons 3 and 4 create and upload the required datasets.

Prior knowledge of entity-relation modeling (ER) or knowledge graph tools like Neo4j, RDF, or OWL maps quickly, but is not required. Lesson 01 lays out the mapping between portal's vocabulary and the familiar one in a single table.

What you can do by the end

• Explain where ontology sits in Portal through the Modeling and Graph Explorer menus under Ontology.
• Define one entity with a valid system name, alias, attributes, Identity Keys, and Display Column.
• Align dataset columns with entity attributes, then load instances through a pipeline Entity output.
• Define a relation between two entities, then load relation instances through a pipeline Relation output.
• In Graph Explorer, select entities and relations, then use Cypher with MATCH, WHERE, RETURN, and LIMIT to inspect a directed relation.
• Know how far a model change propagates (a pipeline's Entity I/O, graph instances, the inspector labels) and update the output contract when a new source dataset arrives.
• Hand analysts a reproducible bundle of collection name, Cypher, and inspection target, and hand engineers the pipeline output contract.

What comes after this Path

Natural directions to go deeper from the ontology flow:

• Build your first dashboard and widget — Put one of your entities onto a widget as input and watch the data architect's model = the analyst's input close inside one portal.
• Your first pipeline — three nodes and one run — Follow the flow of transforming a source dataset and loading pipeline outputs. This is a natural next step when you also own responsibility for populating ontology instances.
• Retail Inventory Intelligence — Every surface in this Path (four entities + three relations) appears tied together in a single domain scenario. The three viewpoints — analyst, engineer, and modeler — meet on top of the same collection. About 90 minutes.

Filling the checkbox next to each lesson auto-records your progress. Go ahead and start.