A strategy for temporal visual analysis of labor accident data

• Autor(a) principal: Brito, Luciana Lima
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/28278; http://doi.org/10.14393/ufu.di.2019.2579
• topic: Labor Accident; Governmental Data; Information Visualization; Acidente de trabalho; Dados Governamentais; Visualização da Informação; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

Labor accidents (LAs) are a serious social problem that can result in physical or/and psychological damages to employees, loss of manpower, expenses with compensations and fines to employers, and social security expenditures with compensations and hospitalizations to the State. The Brazilian Federal Labor Prosecution Office (BFLPO) collects a massive volume of data regarding LAs and this data presents potential to be analyzed. In this work we present a visual strategy employing Information Visualization techniques to explore and analyze LA data, focusing on the temporal aspects to identify patterns on labor accident occurrences, and to associate it to strategical information about how they behave over time. We developed an interactive system to validate our strategy in analyzing LA data from Brazil, and performed the analysis on the data publicly provided by the BFLPO, using the geographical and temporal information associated with the data, to demonstrate the potential of our strategy. As the result of our visual analysis we were able to highlight the evolution of accidents occurrences in the localities of Brazil over time, to identify trends, seasonal events, and abnormal behavior. We could also compare localities' behavior in the same/distinct hierarchical levels, among other task. Our proposed strategy may provide an effective environment to guide the governors in identifying localities lacking attention in order to create public policies to improve inspection, to reduce accidents and to grant safety for employees. The contributions of this work are a visual strategy for temporal analysis of LA data and a computational system, that implements this strategy, to enhance the analysis capabilities of experts from government. We expect to encourage transparency in governments as well as public participation in the governments' decisions.