Geoscience Reference
In-Depth Information
3
Analyzing Spatio-temporal Data: Empirical
and Statistical Approaches
This chapter is dedicated to the presentation of spatio-temporal approaches that
rely on statistical methods (descriptive or explanatory) to analyze changes that occur
over the geographical space and in spatial systems. It is not a manual of spatio-
temporal statistics. The objective is instead to present an approach that relies on
methods commonly used in the social sciences, and just like in the previous
chapters, to discuss the choices made in function of the manner in which space and
time are mobilized. We will therefore assume that the basic statistical methods are
known. We will focus on the approaches that allow, through different types of
combinations and coupling of methods, the highlighting of the main trends of
change, both at the level of the entities involved and that of the relationships that
they maintain as well as the configurations that they draw.
Section 3.1 presents the disciplinary landscape in which the field of spatial
analysis has evolved; the evolution being marked by the statistical methods and the
progress of computer science. This landscape and the interactions between its
various components, including statistics and computer science, have left their
mark on the practices in terms of data processing. These practices are based on a
certain “hybridization” of methods and on the central place accorded to the data
as well as the meaning that is given to them. We then introduce a duality of
objectives for the spatial analysis of spatio-temporal data. This approach consists of
following over time spatial patterns on the one hand and following localized entities
in order to analyze the spatial organization of their evolutions on the other hand.
This duality will serve as a main thread to the developments presented in sections
3.2 and 3.3.
