In this paper we propose an algorithm to identify sets of the most frequently visited tourist sites. We do this by examining the trajectories followed by tourists and by considering their visits to these sites. We propose a second algorithm that recommends a specific order to visit these sites. To accomplish this task, we consider variables such as tourist preferences, departure and arrival locations, and time constraints. To validate our proposal, a prototype website application was developed, which experiments with real vehicle trajectories in Rio de Janeiro. Although more exhaustive experiments will be required to deal with different possible scenarios, preliminary results show the usefulness of our proposal for displaying sets of neighborhoods frequented by vehicles as they move about a city.