Complexity Science for Urban Solutions
Everyone who lives in an urban environment is (consciously or not) affected by its planning and design. Cities have now been recognized as nuclei for innovation, expertise, and opulence; they can be considered as “concentrations of social interactions in space” (Garfield, 2019). As cities grow larger in population and size, they exhibit three key characteristics—complexity, diversity,and intelligence—that offer a glimpse of both the potential of cities and the problems that they face today.
Many urban issues, ranging from wealth inequity to environmental sustainability, are usu- ally tackled independently of each other (Bettencourt and West, 2010) despite their obvious interdependencies. This practice continues the convention of disciplines following a centralized order, which was largely the norm up to the 19th and 20th centuries (Batty and Marshall, 2012). Urban planning in the 20th century was characterized by a rigorous top-down approach, despite notable critics, including Christopher Alexander, who railed against the simplistic urban models of “tree-like” hierarchies and Jane Jacobs, who called for more diversity and citizen-centric design that reflected the realities of urban life. Facing today’s climate emergency, it seems clear that such outdated planning and design strategies are ineffective in satisfactorily addressing many of today’s problems.
The need for cities to become smarter in problem-solving cannot be overstated. It is important that concurrent trends in urbanization, economic growth, technological progress, and environmental sustainability act as drivers in urban planning and design thinking processes. This highlights the need for and potential of artificial intelligence (AI) in urban planning and design today.
Internet of Things (IoT) technology is already ubiquitous in many cities worldwide, with wide-ranging applications in urban planning and design that are based on real-time data collection. In a more dominant role, AI tools and techniques can be tapped into for tackling multiple issues across urban scales, to integrate a conscious top-down approach to planning with site-specific bottom-up solutions. In the following, we detail a complexity science-based methodology that employs machine learning (ML) to quantitatively analyze spaces and activities in high-density urban built environments, with the goal of understanding the efficacy of their use and shortcomings to inform better future planning and design decisions.