From the beginning of the 21st century, cities have become the main expression of spatial organization of society. With population growth, the urban metabolism — all the exchanges between the city and the environment — is modified. We can expect new challenges emerging in the management of, among others, supply sources, energy generation, waste disposal and pollution treatment. In this context, it becomes necessary to understand what will be the environmental impact of a predominantly urban world. Sparse and inconsistent evidences are available, but we are still missing a general and systematic knowledge on the drivers and dynamics of this metabolism. The general goal of this PhD is to understand the evolution of the urban metabolism: how the environmental flows of a city changes as population grows. To achieve this goal, a systematic and programmable method of computing urban metabolisms need to be developed and applied to a diverse set of cities. The results should be analysed using machine-learning techniques to understand their patterns and compare their differences. The resulting findings could bring an overview of the evolution of the urban metabolism, delivering predictive power to managers and indicating systemic sustainable and more efficient ways of governing a city.