Spatial smoothing is a common method to detect spatial patterns or trends, generally represented as hot spots maps. There are not many studies that assess how the different types of Kernel density functions and the bandwidth chosen affect the spatial representation of the data. Consequently, in the study object of this paper we have applied a number of Kernel functions and bandwidths to the same sample of spatial data in order to establish the decisive factor to detect spatial patterns. The results show that, when it comes to detecting spatial patterns, the election of a Kernel function is not as decisive as the election of the right bandwidth, being the latter the factor that more influences in the results.