From Deductive Models to Data-Driven Urban Analytics: A Critical Review of Statistical Methodologies, Big Data, and Network Science in Urban Studies

Urban analytics, which combines geographical analysis, statistics, computer science, and urban planning, has quickly transformed the study and management of cities. The field, which was primarily founded on deductive methods from social physics and location theory, now employs inductive methodologies powered by spatiotemporal big data and machine learning. This comprehensive review traces the evolution of urban analytics from early deterministic models to contemporary network and big data-driven investigations. Special focus is given on the use of statistical techniques for simulation and inference, including the move from equilibrium-based models to dynamic, agent-based, and network models. Recent developments have provided new data sources, such as sensor-generated mobility data and social media, allowing for higher resolution and more comprehensive assessments of urban dynamics. The development of data-intensive approaches has raised serious concerns about privacy, ethics, and the potential amplification of existing societal inequities. This study brings together conceptual advances in network science, the use of big data in mobility and municipal services, and the challenges of combining machine learning with urban ideas. It promotes more critical and emancipatory urban analytics, emphasizing the value of transparency, equity, and the development of comprehensive ethical frameworks in research and practice.