Uma abordagem para a investigação de padrões de movimento e de comportamento de indivíduos empregando análise visual de Predictive Suffix Trees

Abstract

Predictive Suffix Trees (PSTs) are data structures capable of simultaneously represent space, time, and probability. They can be used to predict when a person would leave her current position to move to a new probable location. Although they are usually complex to read, PSTs can help crime investigation; management road traffic; or location-based advertising, for example. This work proposes the application of visual analytics to simplify the task of finding movement patterns and possible behaviors of a person, using data stored in PSTs. For that, we introduce an approach that applies sensemaking and branching time to provide a less abstract character to PSTs, allowing analysts to explore the dynamics of space-time combinations (space, time and space versus time relations) considering probabilities. To validate the proposed solution, we developed a visualization tool and performed three distinct user studies, with a total of 77 participants and two different datasets. The obtained results demonstrated the feasibility of applying the solution, allowing expert and non-experts to solve initial problems, but letting they propose their own questions to find new answers.