IN-PERSON Manchester Livable Cities

Abstract

Smart-city decision-making relies on data-driven systems to allocate resources, design services, and mediate citizen participation. Yet the data infrastructures behind these systems often reproduce sampling, measurement, and historical biases that misrepresent women, migrants, older adults, and people with disabilities. These biases do not merely distort datasets; they shape whose needs are recognized as legitimate, which problems are rendered actionable, and which populations become visible to policy. In this sense, biased data functions as an upstream governance mechanism, structuring participation before citizens ever “engage.” Framed through digital justice, the paper asks how data practices reshape inclusion within urban governance. Conceptualizing the smart city as a socio-technical system, it traces how exclusion emerges across the data lifecycle, from collection and labeling to modeling, deployment, and monitoring/feedback. Using evidence from European smart-city initiatives, it examines how biased data influences priorities and outcomes across key smart-city domains such as mobility, safety, public-space management, and digital participation. Across examples, neutral metrics (usage rates, sensor coverage, incident reports, and app-based inputs) can amplify undercounting and misclassification, producing uneven policy responses and narrowing participation to groups most able to access digital channels and be captured in official datasets. The paper contributes a lifecycle-based analytic framework linking data quality to democratic quality, showing that smart-city “livability” is contingent on whose data counts and whose experiences are excluded. It concludes with mitigation options, including auditing datasets and models for bias, diversifying data collection beyond app users, designing accessible participation channels, validating indicators with community input, and strengthening accountability through participatory oversight.