This thesis presents an implementation of an application used for classifying text data into multiple categories. In collaboration with a company specializing in data analysis in the automotive industry called Skylyze, we developed an application with the purpose of classifying warranty claims. This application consists of 3 main parts - a python backend, an Angular frontend and a Microsoft SQL Server database. With the use of active learning, we were able to create a tool for manually labeling text data that was later used in training a classification model. This thesis first goes over analysis of existing text mining solutions in the automotive industry. Theory regarding natural language processing, machine learning and evaluation methods is discussed in the next few chapters. Further into the thesis we describe used datasets. We also explain the implementation of the application and most significant technologies used for its development. Lastly, we evaluate selected classification models. A comprehensive user manual for the application is attached in the appendix of the thesis.