Modelling spatio-temporal concentration processes of food-retailing groups - a dual kernel density estimation approach (481)

Theme Track: Methods of Spatial Analysis - Spatial Point Pattern Analysis

Authors:
Staufer-Steinnocher, Petra ; Leitner, Michael

This article addresses spatio-temporal changes in the Vienna food-retailing market at the end of the 1990s. By applying the kernel density estimation - an interpolation technique that is appropriate for individual point locations (e.g., retail outlets) - it is possible to spatially quantify market dominance not only for Vienna as a whole but also for the district and even for the sub-district level. The model of the Vienna retail market employs data for which locational attributes are an important source of information. Such data characteristically consist of cross-sections of observations for micro-units such as individuals, households and retail chain outlets at specific points in space, and aggregate spatial entities such as building blocks, census tracts, and store-specific catchment areas. Store location data disaggregated by retail chain, outlet type, average sales area per outlet, and average number of articles sold is available for the fourth quarter of 1998 and second quarter 2001, respectively. These data are represented in form of point pattern maps. Employing the dual kernel density estimation procedure, we analyze the spatial concentration of the predominant retail chains BML, Spar, Löwa, and Meinl in 1998, and BML, Spar, and Löwa in 2001, respectively. The resulting maps indicate areas with increasing and decreasing concentration for each retail chain. A spatially differentiated picture of the market leadership in local micro markets is generated through dual kernel density estimations for the BML against its main competitor, LMS (Löwa, Meinl and Spar) in 1998 and in 2001 (without Meinl).

[Home]    [Sitemap]