Uma abordagem de data mining e business intelligence para otimização do processo produtivo e apoio à tomada de decisão na agroindústria.

Abstract

The evolution and massive use of technologies elucidate the exponential growth of data. This volume of data has been translated into challenges and opportunities for organizations, as extracting relevant information to support decision-making requires changes in processes and actions. However, such behavior changes provide organizations with competitive and sustainable advantages. In this way, the application of data mining techniques combined with business intelligence enables a broad view of the organization, based on the discovery of new models of action, monitoring and improvement of existing processes. In this context, the present study aimed to explore data, through unsupervised learning, to identify implicit information capable of supporting the decision-making process. A systematic literature mapping was developed, considering the appropriation of literary knowledge capable of helping in the phases of the CRISP-DM methodology. As an operation tool, Orange Data Mining was used, as it is expandable and low cost, in addition to supporting a wide variety of data analysis techniques, Google Sheets and Microsoft Power BI, which contributed effectively to computerization and automation of the production process. It was concluded that the achieved results achieved the objectives proposed in the study, presenting relevant information and possibilities for optimizations within the production process. The generated model was capable of transforming the volume of data into important information and the developed dashboard made it possible to easily interpret the results, positively influencing management decisions and generating added and valuable knowledge for the organization.