A Unified Approach to Real-Time Public Transport Data Processing

The use of real operations data is essential for the planning and management of modern public transport systems. With the expansion of universal formats for describing the structure of public transport systems, such as GTFS or Transmodel, the use of these data has expanded far beyond the public transport domain. On the other hand, the effort to use these data encounters the problem of its processing, storage and integration with the structure of the transport system due to the volume and speed of data generation from real operations. These problems are even more evident in the case of further use of these data as inputs for machine learning, or data mining, where integration of data from different systems into a single model is necessary. The purpose of this paper was to design a method by the which big data from real operations could be integrated with the changing structure of the transport system so that this data could be stored long term without loss of granularity, or entropy value. As a result, we proposed a data model with big data transformation algorithm, whose functionality has been verified in testing over the public transport system of the second largest city in the Czech Republic.