Local Measure of Spatial Autocorrelation

While a dataset may reveal a globally significant trend towards clustering, global measures of spatial autocorrelation offer only an ‘average’ and can hide interesting micro-concentrations. To overcome this limitation, local measures of spatial association have been developed. These indicate if one or more confined areas exhibit substantial deviation from spatial randomness. Local measures are particularly useful in large datasets where spatial association between observations is likely to show instability in the form of “local non-stationarity, spatial regimes and spatial drift” (Anselin, 1996: 112).

Local Measure of Spatial Autocorrelation implemented in the AURIN Portal