Remember that astounding map of geotagged photos from London? Well, the data-mining genius who created it has been at work again. Eric Fischer has now added colour-coding to show which areas of town get photographed most by tourists (red), which by locals (blue) and which by either (yellow).
Here's how it works.
Eric pulls out images in Flickr and Picasa that are geotagged (i.e. the photographer has identified the location of the shot through GPS or some other means). He then automatically plots them onto an OpenStreetMap (that's why you can see street lines). 'Locals' are identified by looking at whether the same photographer has taken pictures in London over a range of a month or more. Tourists, by contrast, only upload pics of London once, or over a short time frame. The yellow spots show instances where the data are inconclusive.
Naturally, there are potential flaws in this model (for example, someone who lives in Paris, but hops over to London for regular meetings might be catagorised as a local). But the results seem to reflect what we'd expect. Eric has performed the same experiment on several other cities too. Follow his work over on Flickr.