Why Ontologies Matter for Infrastructure

Every infrastructure decision rests on a model of the world. Whether we acknowledge it or not, the categories we use to describe assets, risks, and communities shape the decisions we make about them.

An ontology — in the data science sense — is a formal representation of knowledge within a domain. It defines the concepts, relationships, and rules that govern how we think about a system. In infrastructure, ontologies determine what gets measured, what gets compared, and ultimately, what gets funded.

The problem with invisible frameworks

Most infrastructure decisions are made using frameworks that are inherited rather than designed. Asset management systems, risk registers, and planning tools embed assumptions about what matters — assumptions that are rarely interrogated.

When a water utility categorises its assets, it is making ontological choices. When a transport authority defines levels of service, it is encoding values into data structures. These choices are consequential, yet they are often treated as purely technical.

Decision-centred data

At Visioning Lab, we advocate for what we call decision-centred data — an approach that begins not with the data we have, but with the decisions we need to make.

This means:

1. Starting with questions, not datasets — What decisions does this data need to support?
2. Making categories visible — Who defined these categories, and what do they privilege?
3. Designing for change — How do we build data systems that can evolve as our understanding deepens?

Towards better infrastructure decisions

The infrastructure challenges of the next decade — climate adaptation, aging networks, equitable access — require us to think differently about data. Not as a neutral resource to be collected, but as a designed artefact that shapes the world it describes.

Ontologies are the lens through which we see our infrastructure. It is time we started designing that lens with the same care we bring to the infrastructure itself.