Abstract
This study applied data analysis techniques to analyze work-related accidents in Brazil's mining sector from 2019 onward, identifying key risks and patterns. Using public datasets from governmental sources, it categorized accidents by the type of injury, causal agents, and affected body parts. The methodology employed included data cleaning, processing, and the development of interactive visualizations using advanced analytical tools, such as Python and Power BI, to facilitate data interpretation. Among the most significant events, the Brumadinho tailings dam collapse in 2019 emerged as a major outlier, substantially affecting multiple aspects of the analysis. This single incident accounted for 71.7% of all work-related fatalities recorded during the four-year period under study, highlighting its disproportionate impact on the dataset. This study also examined the main causes and consequences of mining accidents and facilitated the creation of victim profiles based on gender and age group, incorporating psychological theories regarding risk perception. It was concluded that, although the mining sector represents a small fraction of all work-related accidents in Brazil, the proportion of accidents relative to the number of workers in the sector is substantial, highlighting the need for stricter occupational safety management. The results can guide regulations and help companies and institutions to create safer, more sustainable mining policies. The methodology proved to be highly suitable, indicating its potential for application in safety analysis across other sectors.