The rapidly increasing use of mobile devices for sensing applications, and the growing amount of simulated data generated from simulation tools such as agent-based transport simulations results in massive spatiotemporal datasets. Therefore, the demand for applications to manage and to visualize this data effectively in an interactive and easily understandable way increases quickly. Spatiotemporal visualization has long been recognized as being able to provide a suitable solution for creating such applications. However it remains a challenging task to combine such a visualization with interactive operation. In this paper, we describe MetroBuzz, a prototype system designed for visualizing large amounts of spatiotemporal data, and for interacting with the dataset in a way that is both interesting from a scientific viewpoint as well as for a broader non-expert audience. At its core, MetroBuzz generalizes activities in urban networks, specified as origin-destination trip information in terms of series of line segments in 3D space. These 3D elements are stored in a spatial index that allows quick retrieval of relevant data. Based on this, we implemented interactive tools to define queries on the spatial index in an intuitive manner. We show how such tools can be applied in the case of large transport simulation datasets.