Came across this article in the NYT about "art genome mapping." The idea is to create a massive data base of art through the ages which supposedly can analyze your artistic inclinations and direct you to works of art you may enjoy. Amazon has done this for years with books and Netflix does it with films and Pandora with music. Most internet sites selling things do the same to an extent. For some reason, I feel that the algorithm for predicting taste in visual art may be harder to crack than music and literature. We rarely can say what exactly catches our eye in a painting - the color, the composition, the subject matter, the light, the shadow, the reputation of the artist and sometimes even whether one comes across it in a tasteful museum setting or a roadside flea market. We can to a fair degree of accuracy say for ourselves or about others the *style* of artwork that could have an aesthetic appeal for a particular viewer but not a specific painting. I suppose the Art.sy project will be useful in putting a person in a particular artsy 'box' of sorts but I will wait to see how often its You may also like... recommendation hits the mark with art browsers.
I have written a fair number of times here what I think about formulaic thinking about visual aesthetics. Here is an old post which may be relevant to Art.sy's analytic scope.