Uma Abordagem em Visualização Analítica para Dados Geocodificados de Crimes

Detalhes bibliográficos
Ano de defesa: 2020
Autor(a) principal: Queiroz Neto, José Florencio de
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Tese
Tipo de acesso: Acesso aberto
Idioma: eng
Instituição de defesa: Não Informado pela instituição
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: http://www.repositorio.ufc.br/handle/riufc/53759
Resumo: In recent years, violence has increased considerably in the world. In Ceará, a Brazilian state member, the homicide rate went from 16 per 100,000 inhabitants in 2000 to 37 per 100,000 inhabitants in 2016. With the popularization of spatial databases and Geographic Information Systems (GIS), police departments worldwide started to create various types of crime maps, generated with different techniques, to analyze and fight crime. One of the types of crime maps is the hotspot map, which helps decision-makers to identify high-risk areas and allocate resources more efficiently. The analysis of crime data is usually a complex operation, and a target of Visual Analytics (VA) systems, which are systems that aim to increase human analytical reasoning through visual and interactive interfaces. The need for interactivity requires that VA systems should include high-performance as one of the main conditions. In this thesis, we propose MSKDE - Marching Squares Kernel Density Estimation, a technique to generate fast and accurate hotspot maps. We describe the method and demonstrate its superior qualities through careful comparison with the Kernel Density Estimation (KDE), widely used to generate density maps. Another contribution of this thesis aims to help police departments in their planning activities. Professionals and researchers agree that tracking crime over time and identifying its geographic patterns is vital information for efficient resource planning. To help to perform these activities, frequently, police departments have access to systems that are too complicated and overly technical, leading to modest use at last. We collaborated with domain experts from police departments in Brazil and the United States to recognize and characterize five domain tasks inherent to the activity of tracking crime and resource allocation planning. All domain tasks are related to hotspot analysis and policing, one of the prominent approaches to fight crime. To facilitate the performing of the domain tasks, we proposed SHOC, The One-Shot Comparison Tool, a technique that allows immediate spatial comparison of crime density surfaces. We included SHOC into a VA system, CrimeWatcher, which allows users to perform filtering operations and visualize maps and data smoothly. CrimeWatcher strives for simplicity and will enable users, even without technical knowledge, to perform all tasks, annotate, save, and share analyzes. We also demonstrated that CrimeWatcher and SHOC effectively support the completion of domain tasks in two different real-world case studies.