Analysis of Occupational Accidents in the Metalmechanical Sector in Brazil Using Clustering

Objective: The objective of this study is to use the clustering technique in order to identify patterns, recognize and characterize groups in occupational accident data in the metal-mechanical sector and enable the extraction of information that supports the construction of possible preventive measures. Theoretical Framework: The theoretical framework mainly addresses topics such as work accidents, data mining and the specific task of clustering, in particular the k-means algorithm. Method: The methodology adopted follows the KDD (Knowledge Discovery in Databases) process, with the data mining stage carried out through clustering. Results and Discussion: The algorithm resulted in the formation of 3 distinct groups of data. The main results suggest the predominance of the group composed of workers in the metal products manufacturing industry who suffered cutting and laceration injuries in the region of the fingers, caused by mechanical or accidental agents. Research Implications: The results highlighted the importance of organizations' commitment to promoting the safety of their employees through the provision of personal protective equipment, in addition to adapting their assets in accordance with current safety standards, considering that a large proportion of accidents recorded in the metalworking sector are caused by machines and tools. Originality/Value: This study contributes to the literature by using data mining to seek solutions to Occupational Health and Safety problems. Its application in a sector as comprehensive and representative as metalworking provides great value and support for decision-making by responsible managers and for possible future research.