Utilisation forecast for regional trains Berlin-Brandenburg (RE3 and RE5)

ProTrain wants to create new services or significantly improve existing services by fusion of data and data from other partners provided in the mCLOUD. The aim is to effectively steer passengers before and during the use of public transport in order to make better use of existing capacities and resources in public transport. With the help of the forecasts and the current occupancy levels, passengers should be steered in such a way that the balance between supply and demand is as balanced as possible. For example, a passenger looking for a connection may be advised to use an earlier or later connection to avoid large crowds on the train. For this purpose, requirements analyses were first carried out, which identify relevant user groups and stakeholders with their specific requirements. A comprehensive technical system to meet these requirements was created, a data platform was set up and a modular system landscape was specified, which opened up heterogeneous data sources, and whose data was integrated and provided as a service. Particular attention was paid to the design of efficient data management for historical data. Appropriate interfaces for data exchange between the supplying partners have been defined. For the test user information, a web app was considered the most suitable solution and designed according to the system. With the help of a specially developed SDK on Android phones, the used means of transport were to be determined for some test subjects. These were user-confirmed. The specification of the service has been developed, defining the required data for the applications. Three different forecasting algorithms were developed for specific forecasting periods. The data interface strategy was developed, the components of the interfaces were coordinated between the partners, and the travel archive was made available online. Algorithms for the automatic recognition of means of transport and purpose of travel have been developed. On the basis of passenger counting data and seat capacities, occupancy rates were determined. The occupancy data from AFZS for RE trains were provided to the project. The extension of the timetable data management with weather data has taken place. The baseline forecast was differentiated per weather condition (e.g. summer day). Groups of days, time of day and journey section were calculated and provided. A service concept was created for the forecast information to the test users. A web app on the HaCon platform was developed as a test app. The enclosed data set contains the consolidated forecast based on the three algorithms (column N) as well as an actual load (column O). Columns E and F describe the route from to. The columns J-M provide information about the respective section. The dataset contains all data for a calendar day (November 1, 2019).