If you use one of the newest Google tools, Google Correlate, you can look at typical associations between keywords based on the geographical search pattern.
Here is an example for classical music , looking for debussy and getting liszt :
The list of correlated terms is not really surprising:
0.8860 liszt 0.8473 tchaikovsky 0.8251 g minor 0.8100 c minor 0.8090 tetrachord 0.7887 maurice ravel 0.7856 beethoven 0.7789 well-tempered clavier 0.7774 brahms 0.7760 symphony 5 0.7742 bach 0.7735 prelude in c 0.7689 bach toccata 0.7665 stravinsky 0.7631 bach fugue
This is just a list of very common keywords for students of classical music. A few composers, a few tonalities. The really strange point is why does Google receive a concentration of these keywords from a single state? Why are Utah users so much more interested (or clueless but wanting to know) about Debussy, Liszt or Bach?
Is there something like a geographically pertinent “classical America”, where a significant part of the public is interested in classical music?
this list of correlates to Franz Schubert demonstrates the relevance of this kind of correlation measures in simple (and culturally segregated) cases. All the members of this short list are relevant and could be used for feeding a simple expert system. The first non musical search term is for “diabetes symptoms” later in the list.