1. Trois espaces d’étude comparables et des thématiques susceptibles d’être croisées
Because large cities (especially global cities) are impacted by accelerated dynamics of social, economic and environmental reconfigurations, they make particularly rich case studies. They offer multiple fields of enquiry in terms of both the thematics and the types of data available. Several criteria were used to determine the training sample for this project. An initial sample of three large cities was established based on CIST researchers’ interest for different geographic contexts that are socially and spatially comparable, and on the likelihood of being able to access heterogeneous local data.
The three study areas chosen are Paris, Chicago and Mexico City. The boundaries selected (Fig. 1) do not necessarily correspond perfectly with the perimeters of the three cities’ functional urban areas, which are defined according to centre-peripheries commuting patterns. Île-de- France, for example, do not correspond to the Paris urban area as defined by INSEE in 2010 but rather to the perimeter derived from the general database compiled by the IAU ÎdF (Institut d’Aménagement et d’Urbanisme d’Île-de-France). While the city of Chicago has three official boundaries, we chose to use a fourth, informal perimeter that has been recognised for nearly a century, namely Chicagoland (Chicago Tribune, July 1926), so that the surface area and population of the American city study area is equivalent to that of the Île-de-France region. Finally, although it extends over three states, there is only one official delimitation of the functional urban area of Mexico City available, which is currently recognised by public institutions such as SEDESOL (Secreteria de Desarrollo Social), CONAPO (Consejo Nacional de Población) and INEGI (Instituto Nacional de Estadística y Geografía), and that is Zona Metropolitana del Valle de México (ZMVM). Although less extensive and more densely populated than the other two metropolitan areas, it is the obvious choice.
An analysis of the degree of comparability between these three cities is also dependent on the similarities between the statistical reference grids used. A comparative study of the surface areas, populations and construction methods of the three cities was carried out using reference grids based on an aggregation of contiguous urban blocks. For Paris and Chicago, these clusters of blocks are mainly based on demographic thresholds. In the case of Paris, the IRIS (îlots regroupés pour l’information statistique) generally comprise a population of between 1,800 and 5,000, while the block groups in Chicago cover a population ranging from 600 to 3,000. For Mexico City, the clusters of contiguous blocks (from 1 to 50 blocks) are based mainly on physical demarcations and land cover. Hence, the Mexican urban AGEBs (Áreas Geoestadísticas Básicas) have a larger population range (generally between 1,000 and 9,000) but more homogeneous surface areas than the French IRIS and the American block groups. While their construction methods and definitions differ, these spatial divisions all present similar representation-related issues. The differences in the surface areas between the centres and the outskirts of each of the three cities complicate interpretation and visual comparisons. In the case of Mexico City, the disproportionate surface area of the statistical spatial units is not even visible (INEGI shows only the occasional location of a rural locality). This lack of a common spatial reference for geographic information for the representation of intra-urban data is also problematic for comparative inter-urban analyses. Exploratory studies were conducted to reconfigure a grid that would take into account the entire metropolitan area of Mexico City.
In order to address the challenges relating to the interoperability of conceptually different data, the three study sites are chosen to logically correspond with the thematics selected for cross-tabulation, namely demography, land use and land cover, and the real estate residential market. A growing interest from stakeholders in the new digital data produced by Web 2.0 has also led us to focus on this new generation of local territorial information (for example, the data produced by the ‘community’ platform Airbnb and the social media platform Twitter), which is increasingly being used to analyse urban dynamics, mobility and spatial structures (Louail et al., 2014).
Once we had established the study sites and exploratory thematics, discussion turned to the harmonisation of data, their comparative representation and the way in which they could be interfaced.
