Abstract Due to budget constraints public transportation (PT) can no longer be deployed in regions where it attracts insufficient customers. Novel techniques such as demand-responsive collective transportation (DRT) are evaluated to cut costs. This requires detailed simulations that are able to predict travel demand and include trip execution. Simulating facilities acting as feeder services to time-table based PT services requires detailed and accurate information about the PT infrastructure on a network. However, there are no public data sources that combine network and PT infrastructure data with the preferred level of detail. This led to the development of a new bus stop mapping technique that combines the OpenStreetMap (OSM) and General Transit Feed Specification (GTFS) open data sources, which are maintained independently. Merging the data into a single database requires alignment. Developing bus stop mapping algorithms is challenging due to (i) inaccurate location data, (ii) inconsistent data sources and (iii) the vastly interconnected PT network and services. Due to the inaccuracy in the GTFS stop locations and in the OSM network, pure geometric considerations might lead to multiple candidate solutions to map a stop to the network. The new technique handles all GTFS trips at once and operates under the assumption that PT operators minimize the total distance driven to complete all trips.