Predictive study of the specific properties of the structure of time series of sales volumes of bottled drinking water

Abstract

Importance: this paper presents an expansion and deepening of the previously performed data analysis, consisting of studying sales time series, their systematization and detailed description. The specified tasks are solved at the stage of preliminary or predictive study of time series characterizing daily sales volumes of nine different brands of bottled water. Purpose: аnalysis of specific properties of the structure of time series of sales volumes that determine the dynamics of the key performance indicator for a commercial organization specializing in the sale of one type of goods and, as a result, the selection of relevant mathematical forecasting tools. Research design: the paper presents an expansion and deepening of statistical analysis of sales volume data, consisting in the study of sales time series, their systematization and detailed description. These tasks are solved at the stage of preliminary or predictive research. The study presents calculations of statistical indicators and a vector assessment of time series representing data on sales of a specific brand of water on a certain day of the week. Results: the analysis of the results of a multi-criteria assessment of the persistence or trend stability of time series of drinking water sales decomposed by days of the week and brands presented in the paper is important for obtaining a reliable forecast and developing an effective risk management policy associated with the functioning of a trading company specializing in one type of product.