Developments in the Analysis of Spatial Data

Issues of spatial dependence have arisen in several areas of research, such as the environmental sciences, economics and sociology, but may be more generally relevant in circumstances in which cross-sectional or panel data are collected. Book-length descriptions statistical methods of analyzing spatial data include Cliff and Ord (1981), Anselin (1988), Haining (1990), Cressie (1993), Guyon (1995), Stein (1999), Arbia (2006). The present paper discusses some recent and ongoing developments, mainly from a semiparametric or nonparametric viewpoint. It is helpful for the purposes of this introductory section to fix ideas by discussing a conventional setting of scalar (for simplicity) observations yi, i = 1, . . . , n, having representation