This paper presents a model-based approach to facilitate the development of IoT applications in Transportation Systems. Existing public transportation services are provided by relying on standard data models such as GTFS. However, such models are limited in representing IoT-based infrastructures and the locations that IoT devices cover (e.g., bus seating areas). We introduce a context-aware publish/subscribe IoT platform that supports synchronous data requests, asynchronous notifications and analytics applications. Data requests are created using the system's context, which in our case is based on a transport bus system. Both static and dynamic context properties are modeled by extending the NGSI smart data models. We then introduce a GTFS-to-NGSI mapping tool to enable the enhancement of existing GTFS-based transportation systems with IoT capabilities. We develop a prototype of our platform and we demonstrate the applicability of our approach using open data from the Roma Mobilità bus transportation system.