A Step-by-Step Guide to Building Your Group's Knowledge Graph
Why This Matters
This guide helps you transform your tacit understanding of your research topic into an explicit, structured ontology. Think of it as creating the conceptual architecture of your investigation—the foundation that makes your work communicable, buildable, and inheritable by future cohorts.
The process scaffolds your learning journey:
Understanding — Why do knowledge graphs matter for architectural research?
Operationalisation — How does AT6012 put these ideas into practice?
Application — What does your group's unique knowledge structure look like?
Acknowledgement: This framework builds on the ontology pipeline model presented by Professor Ryan at the LERO Software Research Centre Summit, adapted here for architectural design research contexts.
1Define Your Domain
What it is: A clear, bounded statement of what area of reality you're investigating. Your domain should be specific enough that someone outside your group can understand exactly what you're studying.
DOMAIN: [Specific phenomenon] in [specific context] concerning [specific aspect]
AT6012 Module Example
DOMAIN: Knowledge production and accumulation in architectural
design research education, concerning regenerative
urban systems in Mediterranean port cities
Your Turn
Complete this sentence:
"Our group is investigating _________________________________
in the context of _________________________________,
specifically concerning _________________________________."
Quality Check: Could someone outside your group understand exactly what you're studying? Is it specific enough to be researchable but broad enough to be interesting?
2Identify Core Entities
What they are: The fundamental "things" that exist in your domain—the nouns that populate your research world. These are the categories of objects, concepts, actors, or phenomena you'll be investigating.
List 4-6 core entities for your domain. For each one:
Name it (use CAPS_WITH_UNDERSCORES)
List 3-4 key attributes in curly brackets
Identify 2-4 subtypes
Quality Check: Are these genuinely distinct categories? Could you find real instances of each in your Marseille research?
3Map Relationships
What they are: The connections between entities—the verbs that describe how things in your domain interact. Relationships give your ontology its dynamic character.
Identify 5-8 key relationships in your domain. For each:
Name the relationship (use a verb)
Specify which entity connects to which (use →)
Add conditions or mechanisms in square brackets
Think in terms of: What colonises, produces, enables, transforms, stores, reveals, constrains, or determines what?
Quality Check: Do these relationships capture the dynamics of your research area? Are any relationships bidirectional?
4Articulate Axioms
What they are: The fundamental rules or principles that govern your domain—your theoretical commitments about how things work. These are the claims you're willing to defend.
AX[n]: [Statement of principle or rule]
AT6012 Module Example
AXIOMS:AX1: Knowledge production is collective, not individual—
assessment evaluates group capability, not personal performance
AX2: Each cohort's work inherits from and extends previous cohorts—
knowledge accumulates rather than archives
AX3: Urban form encodes power relations—
spatial transformation is never neutral
AX4: Buildings are dissipative structures within larger energy flows—
non-isolated thermodynamics applies to architecture
AX5: The Living Knowledge Commons requires explicit structure—
ontological framing enables inheritance across temporal boundaries
AX6: Field research and theoretical frameworks must integrate—
neither stands alone
Your Turn
Write 4-6 axioms for your domain. These should be:
Statements you're willing to defend
Principles that guide your analysis
Rules that constrain valid interpretations
They often start with: "When X happens, Y follows..." / "X cannot exist without Y..." / "X is always/never..."
Quality Check: Are these genuinely foundational to your research? Do they differentiate your approach from other possible approaches?
5Synthesise into Knowledge Graph Schema
What it is: Bringing together your domain, entities, relationships, and axioms into a coherent schema that can structure your research outputs.
AT6012 Module Example (Simplified)
# AT6012 Knowledge Graph Schema# Version: 3.1 | December 2025DOMAIN: Regenerative architectural education through
Living Knowledge Commons pedagogy
NODES:
- Research_Groups: G1-G8 with distinct thematic foci
- Knowledge_Artefacts: repositories, dossiers, documentation
- Theoretical_Frameworks: assembly, thermodynamic, post-human
- Field_Sites: Marseille neighbourhoods, Irish comparisons
- Contributors: guest lecturers, previous cohorts, future cohorts
EDGES:
- PRODUCES: group → artefact
- APPLIES: group → framework
- DOCUMENTS: artefact → site
- INHERITS: current_work → previous_work
- TRANSMITS: current_work → future_cohorts
- CONNECTS: module → european_networks
CONSTRAINTS:
- All artefacts must be CC BY-SA 4.0 licensed
- Assessment is group-based, not individual
- Knowledge must be structured for inheritance
- Theoretical frameworks must connect to field evidence
Your Turn
Create a similar schema for your group's research domain, bringing together all the elements you've developed in Steps 1-4.
Common Pitfalls to Avoid
Too Abstract: "Urban systems" is too vague. "Informal water infrastructure in northern Marseille communes" is specific.
Too Many Entities: Start with 4-6 core entities. You can always add more later as your research develops.
Relationships Without Directionality: Always specify which entity acts on which. "X relates to Y" is weak; "X transforms Y" is strong.
Axioms That Are Actually Observations: "There is dereliction in Marseille" is an observation. "Dereliction enables alternative actor-networks to emerge" is an axiom.
Forgetting the "Why": Your ontology should help you do better research. If it doesn't clarify your thinking, revise it.
Next Steps
After completing this exercise:
Document your ontology in your group's repository section
Visualise your knowledge graph as a diagram (we'll cover tools for this)
Apply your ontology to structure your Marseille field documentation
Refine as your research develops—ontologies evolve with understanding