On this run I used a large value of epsilon (2Km); the meaning of that value is that we accept large clusters (up to 2Km); since the algorithm “merges” clusters, we will end up with some very large clusters, that will have almost certainly a low density.