• spare part management
• text mining
• maintenance
• data mining

This thesis was performed in collaboration with the R&D team of LogObject AG, partly carried out in the Zurich site. The main goal of the project is the development of a system able to predict the specific type of materials and spare parts necessary for the maintenance of household appliances, produced by a LogObject customer. The data, analysed with Python, was extracted from the mLogistics software, developed by LogObject and used by the customer to manage the activities of the operators in maintenance tasks. During the project I achieved the following activities: acquisition and exploration of data extracted from the software; development of three main predictive models based on text mining, machine learning and frequent patterns mining; comparison of different models and identification of the best performing one. The analyses led to the creation of a versatile system with wide application possibilities for other customers of the company, able to increase the efficiency in maintenance activities. In addition, the basis for possible future improvements of the project has been laid, starting with the proposal for a new data collection method.