Understanding the dynamics of vehicles is a major issue in smart cities evolution. Most intelligent transportation systems concentrate on macro- mobility with centralized decisions and global optimizations for whole cities. With the great development of autonomous cars and vehicle-to-vehicule communication systems, micro-mobility is again of special importance for opportunistic data dissemination and local decisions - for instance when arriving at traffic lights or when vehicles meet. Simulation tools are widely used to evaluate assumption correctness and performances of optimization algorithms. Several macroscopic vehicular mobility trace exist, but few detailed ones at the microscopic level. Most of macro traces integrate simplistic models for intersections and roundabout, while the complexity of this fine-grained mobility can greatly affect the global optimization. For that purpose, we propose a dataset describing a complex roundabout in Creteil, France.