Facilities location is considered by many researchers and professionals to be an important strategic decision within the design of the supply chain network. Location problems are characterized from a space which the distance is defined, locating a set of new points in a fixed or variable number in order to minimize the cost of satisfying the demands given with respect to some set of constraints. This dissertation presents the use of the driving distance for selection of locations based on calculated demand for products with controlled temperature into Brazilian territory. This analysis is based on k-means algorithm for grouping and suggestion of the best location, respecting the distance and consumption requirements. As proof of the concept, the computational environment, the source code in R, the calculation of demand on a municipal scale, the relationship between market segmentation and the number of clusters are presented. We believe that the method presented in this dissertation can be applied to improve the management of supply chain decisions and help companies to access facility locations more quickly.