The Uses of Big Data in Cities

There is much enthusiasm currently about the possibilities created by new and more extensive sources of data to better understand and manage cities. Here, I explore how big data can be useful in urban planning by formalizing the planning process as a general computational problem. I show that, under general conditions, new sources of data coordinated with urban policy can be applied following fundamental principles of engineering to achieve new solutions to important age-old urban problems. I also show that comprehensive urban planning is computationally intractable (i.e., practically impossible) in large cities, regardless of the amounts of data available. This dilemma between the need for planning and coordination and its impossibility in detail is resolved by the recognition that cities are first and foremost self-organizing social networks embedded in space and enabled by urban infrastructure and services. As such, the primary role of big data in cities is to facilitate information flows and mechanisms of learning and coordination by heterogeneous individuals. However, processes of self-organization in cities, as well as of service improvement and expansion, must rely on general principles that enforce necessary conditions for cities to operate and evolve. Such ideas are the core of a developing scientific theory of cities, which is itself enabled by the growing availability of quantitative data on thousands of cities worldwide, across different geographies and levels of development. These three uses of data and information technologies in cities constitute then the necessary pillars for more successful urban policy and management that encourages, and does not stifle, the fundamental role of cities as engines of development and innovation in human societies.