This study addresses a spatial cluster analysis by multivariate statistics to investigate the relation between soybean yield and the following meteorological variables: precipitation, average air temperature, solar radiation and Moran's local index (LISA). The study was conducted with harvesting data from the crop calendars from 2000/2001 to 2007/2008 in the Western region of the State of Parana. The identification of the appropriate number of clusters for each crop year was obtained using the minimization of deviations. The study showed that it is possible to form groups of municipalities using the similarities of the variables under consideration. Cluster analysis is a useful tool for better management of production activities of agriculture, according to which the grouping is possible to establish similarities that provide parameters for better management of production processes that bring both quantitatively and qualitatively, results sought by the farmer.
Spatial statistics area; spatial similarity; multivariate analysis