Synthetic Dataset for Similarity: Dataset generated with the Hermoupolis semantic trajectory generator, that allows creating trajectories based on pre-defined profiles. It generates semantic trajectories where both stops and moves are semantically enriched with annotations defined by the user. Therefore, it is possible to enrich the moves between the stops with semantic information, such as the transportation mean, the activity performed during the move, the name of the streets, etc. Hermoupolis has several parameters to simulate real trajectories, such as the definition of the average time of the moving object at each stop, the standard deviation of the time of each stop, the speed of the moves, sampling rate, and so on. We generated 440 trajectories with several stops and moves, using as semantics of the stops the POI category and the activity performed at the stop. For the moves we generated the raw points with the following attributes:(i) the transportation mode; (ii) the activity performed during the move; (iii) the travelled distance; (iv) the average speed; and (v) the duration of the move.