Ontologías: cómo encontrar significado en los datos

What is an ontology?

In today's digital ecosystem, where data is abundant but meaning is scarce, a fundamental question often goes unasked: what do these data actually mean? While traditional questions about data have focused on operational aspects — where does it come from? where is it going? — the more important question concerns interpretation and meaning. Ontologies provide the answer.

The concept of ontology has its roots in classical philosophy. From the Greek οντος (relating to being or what exists) and λόγος (science, study, theory), it constitutes that part of metaphysics concerned with what there is. Aristotle called this branch of knowledge "Metaphysics", establishing what he termed "First Philosophy" or "Ontology": the study of what exists, is real or has being. Ontology is concerned not only with entities, but also with the way in which they relate to one another.

This age-old philosophical concept has found practical application in computer and information science over recent decades. In this field, an ontology is defined as a structured set of definitions expressed in formal language that specifies the classes, types, attributes, properties and relationships between entities operating within a given domain of reality and knowledge. What the philosophical and computational approaches share, despite their differences in context and purpose, is the attempt to systematically map data to meaningful semantic concepts.

Ontologies are represented visually with boxes (real-world entities) and arrows (relationships) to aid human interpretation, but they also support machine-readable representation through languages such as OWL. This duality is crucial: it allows humans and computational systems to "converse" and work within a shared framework of meaning. Although data may influence their structure, an ontology must function independently of the specific content of the data in a given ecosystem. Its purpose is to define what makes sense, providing a common framework that connects data with their meaning.

In relation to information systems, ontologies function as their "cognitive artefact" or "artificial mind" for two fundamental reasons. First, their system of classes, attributes and relationships corresponds to the structure of reality for a given knowledge domain. Second, and more importantly, the logic they contain — whether first-order logic or predicate logic — enables cognitive operation over that reality: that is, automatic reasoning and the extraction of implicit knowledge from the defined relationships. This unique advantage makes it possible to tackle problems in any field where decisions cannot tolerate the margin of error inherent in probabilistic reasoning.

Why does an ontology matter?

Understanding what an ontology is is not enough without understanding why adopting one represents a significant strategic advantage. Ontologies make it possible to establish a common vocabulary among the actors and systems operating within a shared environment, facilitating collaboration, mutual understanding and integrated workflows. By standardising domain semantics, they also make underlying conceptual assumptions explicit, contributing to better decision-making.

Their use in information systems allows relevant data to be automatically mapped to conceptual classes, enabling applications and processes to "know" how to interpret that data. As a result, the need for custom programming is dramatically reduced and the development of ontology-aware applications is made easier — applications that cease to be mere processors and become intelligent interfaces that enhance operational efficiency.

An effective ontology acts as connective tissue between data and applications, fundamentally transforming the way information systems are developed and operated. Data integration becomes a matter of mapping raw data to an ontology. Application development becomes a matter of creating ways to interact with ontological objects. Standardised logic can be embedded in the ontology itself to ensure consistency across applications.

When applied to a set of concrete facts, the ontological model makes it possible to build a knowledge graph: a structured representation in which entities, types and relationships are expressed through nodes and edges. By describing the logic and structure of a knowledge domain, the ontology acts as the foundation that allows the knowledge graph to capture, relate and contextualise the data it contains.

One of the key characteristics of ontologies is that, by integrating the essential relationships between concepts, they enable automated reasoning over data. This reasoning is straightforward to implement in semantic graph databases that use ontologies as semantic schemas. Beyond mere static representation, ontologies function as a "brain" that works and reasons with concepts and relationships in ways similar to how humans perceive interrelated concepts.

In addition to this reasoning function, ontologies provide more coherent and intuitive navigation as users move from one concept to another within the ontological structure. They are easy to extend, as relationships and concept mappings can readily be added to existing ontologies, allowing the model to evolve as data grows without affecting dependent processes and systems. They also provide the means to represent any data format — including unstructured, semi-structured or structured data — enabling smoother data integration, simpler concept and text mining, and data-driven analysis.

As one of the fundamental pillars of semantic technology, ontologies form part of the W3C standards stack for the semantic web, enabling data to be linked together in the Linked Data Web and providing a common structure that facilitates interoperability between distributed and heterogeneous systems and databases.

Other formal approaches to knowledge representation exist — such as vocabularies, taxonomies or thesauri — but unlike these, ontologies do not merely categorise: they also express complex, multiple relationships between concepts, offering greater semantic richness and linking capability.

Domain ontologies

A domain ontology represents concepts and relationships belonging to a specific part of the world, and as such enables highly specialised knowledge to be managed. In the field of information science and technology, where there is a tendency to restrict and control vocabularies as tightly as possible, these domain-specific ontologies are the logical outcome of efforts to formalise and compute the content of digital resources.

In this type of ontology, the meaning of each term is defined by the context of the domain in which it is applied, enabling a precise and operational representation of that knowledge. However, given their focused nature, domain ontologies tend to be external to one another — that is, they do not share a common semantic foundation — which limits their direct interoperability. This limitation poses a challenge when the reality being modelled does not respect the artificial boundaries of domains.

Ontological hybridisation

Reality rarely allows itself to be neatly compartmentalised. Knowledge domains, far from being pure or isolated, tend to intertwine and overlap, reflecting the complexity and continuity of the world we are trying to model. In practice, real organisations and systems require more flexible models: hybrid ontologies or expanded ontologies, capable of integrating different perspectives into a single coherent representation.

This type of modelling requires the creation of a higher-order ontological narrative, built from controlled vocabularies developed across different cultural, scientific and linguistic traditions. Ontological hybridisation is therefore a delicate, craft-like process that responds to the need to computationally represent areas of reality that exceed the capabilities of pure domain ontologies.

Use cases

The case for adopting ontologies should by now be fairly clear. By precisely defining the terms, relationships and rules that structure an area of knowledge, ontologies become reusable semantic infrastructures. They are capable of capturing the implicit meaning of data, guaranteeing interoperability between systems and enhancing knowledge management — making it possible to connect scattered data, infer new relationships and automate processes based on the logic of the represented domain.

Beyond these immediate benefits, ontologies allow a data ecosystem to grow and evolve sustainably, preventing increases in volume or complexity from leading to a loss of control or semantic inconsistencies. Far from being limited to representing static knowledge, they provide a dynamic structural foundation capable of adapting to new sources, contexts and needs without compromising the integrity of the system.

In practice, ontologies are applied in contexts such as finance, healthcare, law, defence, cybersecurity, cultural heritage management, education and public administration, demonstrating their versatility and strategic value in domains where precise knowledge and consistent interpretation are essential.