2. Ontology Development

"What are you, in
Love with your problems?
I think you take it
A little too far..."
- Talking Heads "No Compassion" [Album: Talking Heads 77] 

A word on Ontologies

An ontology-driven approach is a key success factor for Knowledge Graph projects. Ontology development can also be a quagmire from which your project may never escape. It is easy to “take it a little too far” and lose site of the goal of answering the Competency Questions by attempting to model all possible concepts and relationships.

In her very approachable book A Data Engineer’s Guide to Semantic Modeling, Ilaria Maresi describes ontologies as part:

• Dictionary - containing terms and definitions
• Taxonomy - class hierarchy
• Thesaurus - relationships between terms

To which I would add:

• Rules and Restrictions - rule-based group membership, inclusions, exclusions, and type definitions

An ontology provides the capability of inferring values and relationships not in the original data. While much has been made of this capability, there are many pitfalls to be avoided. Only basic ontology examples are illustrated in this project. Please see the Resources page for information about ontology design practices and readings that provide in-depth coverage of the power of ontologies.

With these cautions in mind, it is time to start defining the data model and relationships.