Estudo de caso em um veículo publicitário com dados de contagem longitudinais

Abstract

After World War II, the interest in analyzing populations over time simultaneously increased with technological advances, facilitating the use and analysis of large databases. In this type of study, the focus is to see if a certain characteristic or behavior of a study group changes over a given period of time. Such technique is called Longitudinal Data Analysis. Along with the Longitudinal Data Analysis, we introduce concepts of Generalized Linear Models (GLMs), which are very useful statistical tools for modeling a particular variable of interest. Through GLMs, we can use different information obtained from a database allowing us to define an equation that represents the relationship between independent variables and a variable of interest in a single event. This monograph presents a case study which considers longitudinal count data, aiming at verifying the influence of a particular trait in the use of promotions contained in an advertising vehicle and also check if the volume of uses varies over time. For this, we analyzed the behavior of users of this advertising vehicle for a period of 12 issues and applied Poisson regression methods for longitudinal count data in order to describe the profile of users and to assist the decision making process in the company where data was obtained.