|
From: Paul H. <pa...@on...> - 2010-02-23 19:17:49
|
Hello, We just launched a new site at http://ny-pictures.com/nyc/photo/ which is based on data from dbpedia and freebase and uses openstreetmaps for mapping. Behind it all is a 'semantic GIS' engine that combines the ability to represent traditional GIS with the ability to make assertions such as "The Empire State Building is in Manhattan." A particularly remarkable feature is that very few of the images are geotagged: we're able to establish the locations of the photographs based on text and other available evidence. We've currently got an internal RDF vocabulary that does a good job of making assertions like "Picture A is a picture of topic B" "Image C is an instance of Picture A" "Image C is 473 pixels high" "Image C is available at <http://...>" We're thinking a lot about how we're going to expose these assertions to the general public: we used SIOC for car pictures, but it doesn't seem to be all that useful for the major use case I'm thinking of: providing end users with the ability to query an endpoint for creative commons images about topic X and get back a list of images [in different sizes] with sufficient metadata to use usefully. Any suggestions or ideas? |