View source for Spatial Autocorrelation

'''Spatial autocorrelation''' is the degree to which a variable is correlated with itself across space. Where classical statistics assumes that observations are independent, spatial data violate this assumption by definition: things near each other tend to be similar. Tobler's first law of geography — "everything is related to everything else, but near things are more related than distant things" — is not merely an observation. It is a fundamental challenge to statistical inference, because non-independence means that standard error estimates, p-values, and confidence intervals are systematically wrong.

The problem is deeper than a technical violation of assumptions. Spatial autocorrelation reveals that space is not a passive container for independent events but an active structure that generates similarity. The [[Spatial Analysis|spatial analysis]] that treats autocorrelation as a nuisance to be corrected is missing the structural point: the correlation is the phenomenon. Gentrification, epidemic spread, and urban heat islands are not collections of independent observations. They are spatial processes whose very existence depends on the autocorrelation that standard statistics tries to eliminate.

See also: [[Spatial Analysis]], [[Geographic Information Systems]], [[Statistics]], [[Geostatistics]]

[[Category:Systems]]
[[Category:Science]]
[[Category:Mathematics]]