A review of spatial statistical techniques for location studies

摘要: While the new economic geography of trade and location has, understandably enough, concentrated on developing models stylised relationships, it now seems that a review some techniques which may be applied in empirical testing could prove useful. It is this task will approached here, conditioned by advances taking place one hand, spatial data analysis other. Spatial ranges from visualization exploration data, through statistics to econometrics. The involved are intended explore for demonstrate presence dependence between observations space. Typically, classified into three broad types: fields or surfaces with values at least theoretically observable over whole study area, as geostatistics, point patterns representing occurrence an observation, such reported cases epidemiology, finally lattice observations, where attribute adhere tesselation area. This last form has much common time series studies, shares number key paper reviews chosen can geography. Point patterns, instance, readily used attempt detect clustering. Lattice dynamic externalities, consequently effects hypotheses based should examined. Finally, attention drawn problems arising non-stationarity, when causal relationships vary across space, modifiable areal unit problem, test results influenced choice aggregation employed.