Linking VGI for place-based map generalization

Traditional map production is a top-down process driven by national mapping agencies. We argue that VGI have the potential to allow us to create place-based maps, and in particular that VGI in the form of Flickr data can be used to drive generalization operators applied to Open Street Map data. We introduce a method to link the descriptions associated with georeferenced Flickr images to Open Street Map geometries for popular London locations, and demonstrate that rich semantic data, contributed by many users, are available after filtering for bias. We illustrate the potential application of our approach to three generalization operators