Data collection and analysis processes are increasingly important for both small businesses and large global enterprises. Data analytics are essential for process control and for providing insights for strategy development. Business intelligence (BI) analysis points out opportunities arising from large volumes of data and is the basis for assertive choices. The objective of this study was to conduct Big Data Analytics to develop a predictive evaluation of machine failures in production in order to propose an effective preventive maintenance plan. The research method adopted was to perform Big Data Analytics to develop descriptive and multivariate statistical evaluation in Python through the investigation of 10,000 occurrences referring to production process records with three types of products and quality variants. The result revealed information about the main failures and their causes, pointing out that the main problem in production is tool wear, clearly caused by excess air and process temperature. Using multivariate statistical analyses, we demonstrated the relationship between process temperature and air temperature, as well as the adverse relationship between rotational speed and torque, and concluded that there were also significant failures from overexertion and lack of energy. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.