Computational Social Science (BSc, Hons) @ University College Dublin

_{(Updated on 12/02/2026, Initially Written on 05/01/2026)}

What is Computational Social Science

Computational Social Science (CSS) is a field concerned with understanding how social systems behave when they are treated as data generating processes. It’s multidisciplinary, sitting at the intersection of sociology, economics, statistics, mathematics, and computer science, but it is not a simple blend of those disciplines. Its core aim is not prediction for its own sake, but responsible modelling using data and computation to analyse power, incentives, institutions, and behaviour, while remaining explicit about assumptions, limits, and bias.

CSS treats models as arguments with data being something produced by social and institutional systems rather than neutral facts, and computation as a tool that is always embedded in political and social context. The emphasis is on explanation, accountability, and interpretability, especially in domains where models influence real decisions.

My Training Within CSS

Within this framework, my own training has been anchored in sociology and economics, with computation and statistics serving those lenses rather than replacing them. Sociological theory grounds my understanding of power, inequality, race, migration, surveillance, and institutions, while economics provides structured models of incentives, markets, labour, growth, games, and policy trade offs.

Computation, statistics, and mathematics were treated as enabling tools rather than ends in themselves. This meant learning to code, model, and analyse, not to optimise benchmarks, but to test claims, stress assumptions, and make social systems legible without stripping away their complexity. As a result, I’d say that my undergrad academic profile is best described as applied economic and sociological research with strong computational fluency, rather than generic data science or pure theory.

In actual practicality, this means working with large, messy datasets, building transparent econometric and computational models, testing institutional or policy claims against data, and communicating results clearly to non technical or policy stakeholders.

What This Trained Me to Do - Explicitly - in bullet points

• Frame social and economic questions precisely - for effective modelling.
• Choose appropriate methods (statistical, computational, or theoretical) rather than defaulting to a single tool.
• Work with messy, biased, real world data and document its limitations.
• Interpret quantitative results through institutional and historical context.
• Communicate technical findings clearly to technical and nontechnical audiences alike.

Where My Interests Now Sit

Over the course of the programme, my interests gradually settled and became more concrete. I found myself caring less about modelling techniques in isolation and more about what happens between agents, and how those interactions scale into groups and institutions.

Learning across econometrics, labour markets, growth, surveillance, migration, and computational modelling painted a clear picture: intelligence rarely stays individual. Beliefs about other people, relationships, reputations, roles and institutional rules change what an agent can do and how others respond.

My current frontier is therefore social intelligence: how artificial agents model other minds, maintain relationships across time, coordinate in groups and eventually participate in larger artificial societies. This makes beliefs and relationships computational objects without pretending they are neutral facts.

The larger ambition is to build inspectable synthetic social worlds in which claims about memory, coordination, norms and institutions can be tested rather than merely narrated. Earlier work in AI safety and governance still matters here, but as one source of institutional questions rather than the identity of the programme.