Distribuição espacial e temporal de vendas de pequenos e médios varejistas do estado da Bahia

Abstract

Statistics is a very useful science in the interpretation of data and aid in decision making. Geostatistics, for example, is a tool used to study the spatial variability. It allows the interpretation of the results based on the structure of the natural variability of the evaluated attributes, and considers the spatial dependence within the sampling interval, since the time series analysis models the temporal behavior of the variable and the model used to make projections. The objective of this work was to use the geostatistic analysis to make a mapping of the average sales based on the past 12 months 171 customers of Bahia and hold a sales projection to 10/2016 based on the information of the last five years. For this, the databases used in this study were provided by a wholesale company of Uberlândia. The procedure for analysis was conducted in three phases of analysis: descriptive, geostatistics and time series. The software used for analyzes were geoR, SPSS, SAS Visual Analytics and GS +. In geostatistic analysis it was found that for the analyzed variable was no spatial dependence. In the temporal analysis were adjusted several models to explain the series and the best fit models were SARIMA ( 0,1,1 ) ( 1,1,0 ) and Holt -Winters additive, the latter being used for the projections. Statistical operating data of the companies are very important tools in planning and making decisions them.