The Internet of Things (IoT) plays an important role in the development of smart cities. In this paper we focus on the development of IoT-based smart services for solving urban problems that involve IoT-enabled Observation, Orientation, Decision, and Action (OODA) loops. We also focus on how to efficiently support such OODA loops in situations where such loops involve internet-scale data. More specifically, IoT supports Observation via the discovery of sensors and the integration of their data. It supports Orientation via a contextualisation process that refines such data to include only those that are relevant to the situation and/or activities of each specific individual or group. As IoT contextualisation potentially involves internet-scale data, performing this process efficiently allows for fast decision making, and this in turn permits carrying out a timely Action. In this paper we propose an approach and related techniques for performing internet-scale data contextualisation. In particular, we propose IoT-based contextualisation techniques that effectively consider the entire range of data that is being collected in smart cities and use such data to provide hyper-personalised information to each user, i.e., information that best suits the context of each user in the Smart City. We exemplify the proposed contextualisation solution in a smart parking space recommender application/service, and provide an experimental evaluation of this service to illustrate the benefits of our solution.