###Towards a New Science of Urban Form
Project ID: 2228bd1143 (You will need this ID for your application)
Research Theme: Information and Communication Technologies
UCL Lead department: Computer Science
Lead Supervisor: Licia Capra
Project Summary:
Back in 2016, the United Nations recognised that “all people [should] have equal rights and access to the benefits and opportunities that cities can offer, and the international community [should] reconsider the urban systems and physical form of our urban spaces to achieve this”. Despite the relevance for society at large, the impact of urban form on cities’ social, economic, and environmental functions remains mostly anecdotal or speculative to scholars, hence undecipherable to policy and practice. One explanation is that we still do not have a solid and comprehensive method to describe, classify, and evaluate urban form. To address this gap, research at the intersection of computer science and architecture (urban morphology in particular) is needed to explore a fundamental question: “are there recognisable patterns of city form that contribute towards their performance?” In this PhD project, we will start unpacking this question. Starting from freely available satellite imagery, we will develop AI techniques to automatically extract metrics of urban form (e.g., street network, building footprints, digital terrain mode, etc.) so to offer a numerical, comprehensive, and granular characterisation of urban form. Agglomerative hierarchical clustering algorithms will be investigated to learn “urban types” from these raw urban form metrics. Finally, novel models of spatial data analysis, inspired by quantitative geography and spatial econometrics, will be investigated and advanced with techniques for reasoning in the presence of uncertainty and missing data, so elicit the relationship between urban types and key performance indicators of prosperity, wellness, resilience, and climate/energy functions of places.