Bridging the Gap between Linguists and Technology Developers: Large-Scale, Sociolinguistic Annotation for Dialect and Speaker Recognition