Tantoush, M., Sengupta, U., Solomou, S., Chueng, E., 2024.

Bridging Urban Big Data and Agent-based Modelling: A New Theoretical Framing for Uncovering Complex Patterns and Mechanisms

Urban environments are inherently complex, characterized by numerous interconnected components and
dynamic interactions. However, there is a lack of clear theoretical framing around sensing this complexity
to understand spatial and temporal interaction patterns that emerge through self-organization and
underlying mechanisms. The convergence of urban big data and complexity theory provides a novel lens
through which the intricacies of urban systems can be studied through patterns. This paper explores a
new theoretical framework situating complexity theory as bridge between analysis of urban big data as
emergent patterns and Agent-based modelling self-organising mechanisms. A critical comparison is made
between bottom-up (agent-based modelling) and top-down approaches for uncovering emergent patterns
(urban big data). Top-down urban big data analysis typically focuses on detecting patterns and correlations
at the empirical level. In contrast, agent-based modelling simulates the interactions of individual agents
to explore self-organisation rules towards uncovering mechanisms at the real level. These considerations
are discussed and a new theoretical framing proposed to overcome the current gap between the two
different approaches and explore their synergies. We argue that integrating these approaches, simulation
mechanisms-based approach to analyse and understand emergent patterns, can provide a new way to
study urban complexity by linking patterns with causal mechanisms.