ARCHITECTURE OF THE PASSENGER CLUSTERING SYSTEM

Background. The paper raises the problem of building a software architecture for the primary processing of airline passenger data, data structuring and further clustering, taking into account industry specifics. Materials and methods. To solve this problem, the best practices of building loaded systems for working with big data were studied, the most promising from the point of view of development and mutual integration were identified. Results. The architecture of the passenger clustering system chosen as a result of the work was successfully implemented in practice and proved to be effective and reliable. Conclusions. The described approach is recommended for use as well-proven in working with large arrays of industrial data. The approach will allow flexible scaling of the data processing system and connection of other modules to it in the future.