Our project’s aim is to produce data bases for international comparison, where data are harmonised in space and time according to explicit conceptual, statistical and computer models. Their multiple and flexible conception enable a variety of hypothesis testing. We propose an integrated, documented and reproducible method that is illustrated by different case studies representative of the world-wide diversity of urban systems (France, United States, South Africa…).
We produce a model-assisted corpus of comparable data. Our method consists in complementing the existing historical data bases by new sources of information like Corine LandCover. This system is adapted to the management of heterogeneous and partial information. It enables flexibility in the production of urban data. This information is fully documented, from sources to metadata. We include three levels: the elementary spatial units which compose an urban entity, the cities themselves (defined either as urban agglomerations or functional urban areas), and the systems of cities (within regional, national or continental limits, to be discussed explicitly).
The metadata are made explicit by providing a series of analytical tools for measuring urban growth, evaluating urban hierarchies, characterising urban functions and competences and modelling their dynamics. The data on urban populations will also be accessible on this website. Different works will demonstrate the impact of data measurement and definition on the evaluation of the heuristic and predictive use of usual models of urban dynamics. All documents on data base construction, sources and methods, soon will be put in a website with interactive interfaces.
The estimations of the city sizes is controlled and validated through a variety of conceptual models of cities and urban systems, as well as models of urban growth and urban hierarchies. We use as possible a time definition of cities (the one-hour time-budget), that allows a choice between municipality, agglomeration or functional area according to the time period, the national or the hierarchical context. The sensitivity of these models to the definition and delineation of urban entities is tested in a systematic way. It contributes to evaluate the relevance of these models, which is now controversial in urban literature, and we expect to demonstrate the enormous incidence of the quality of data on the results.