This question has been rejected from both Stack Overflow and the new beta site Open Data. It was suggested on the Open Data meta site that it would be more appropriate for Linguistics. I'm not convinced of this, but I'm willing to at least ask. In any case, there is a "corpora" tag.
(There's also a discussion on Meta Stack Overflow about its appropriateness there: https://meta.stackexchange.com/q/179125/218334)
EDIT: The above link won't work for anyone who isn't a private beta member for another day or so, apparently, so here's the text of the question:
I'm interested in doing sentiment analysis in languages other than English; for instance, I would like to analyze tweets written in Spanish. However, it appears that sentiment dictionaries for non-English languages either do not exist at all or are not comparable to the sentiment dictionaries available for English.
How can I go about constructing such a dictionary? How useful, for instance, would a standard Spanish dictionary be?