Mining tourist routes using flickr traces

This paper is about a new methodology to automatically rebuild main paths from Flickr´s traces. Our geotagged image metadata´s corpus allows the construction of each photographer´s timeline. The tourist´s itinerary in the destination can then be reconstructed considering two modes: the fastest path and the most likely path. Major paths are directly extracted from the fusion of all itineraries. To illustrate the multi-scale efficiency of this work, several fields of study are presented : Berlin, the Loire Valley and the Palace of Versailles.