Identify challenges in spatial statistics, including potential limitations of area-unit based observations, computational limitations and uncertainty
Spatial statistics is dedicated to describing and modeling georeferenced data through the application of statistical theories and methods. Unlike conventional statistical approaches, which often assume independence among observations, spatial statistical techniques allow to account for locational aspects observations in addition to their attributes. Modeling georeferenced data with conventional non-spatial statistical approaches can lead to bias and unreliable results. This article first discusses measurements of spatial arrangements including mean center and standard distance deviation. It then reviews statistical methods for the types of spatial data—point data, geostatistical data, and areal data. Following this, it examines Bayesian spatial models, which offer a flexible framework for incorporating spatial dependence. Finally, the article concludes with a discussion of ongoing challenges in spatial statistics, including potential limitations of area-unit based observations, computational limitations, and issues related to data uncertainty.