Cool idea, but I think the data is skewed by population density, for example, the map shows Juarez as being more touristy than the Grand Canyon. A better approach might be to map the ratio of photos to population density.
Many places. World Bank, Census, IMF etc. Do a text search on this page: http://uuorld.com/portal/ for "set:" Each set is free from its source but they require lots of tedious massaging to reformat.
Agreed, it pretty much follows population density. Some notable exceptions are Utah, Colorado, Yellowstone, the Grand Canyon, the Smoky Mountains (on the border of Tennessee and North Carolina), the Carolina beaches, Vermont, northern New Hampshire, and the Adirondacks of New York.
Another exception is the Oregon Coast (where I grow up) It's bright yellow but pretty much no one lives there (relatively speaking). For example, the county I grew up, Lincoln County, is roughly the size of Rhode Island but has only around 45,000 people (figures from memory).
But it is extremely touristy. Highway 101 grinds to a stand still most summer days, especially when it's really hot inland in Portland / Salem / Eugene.
Upper Michigan has some definite breaks with population density. Escanaba has nothing, while Mackinac Island is very bright. Closer to home, there are hotspots at Houghton/Hancock and Copper Harbor, but emptiness in between, whereas Calumet/Laurium is quite a bit more populous than anywhere farther up the peninsula.
I am the author of the map. Sure, I will take population into account in the next revision. My main focus when creating this was to discover less touristy places, so I did not pay much attention to cities.