Table 1.

Training data for the word-embedding models.

Rural-urban commuting area tier	All tweets	Urban core	Small town/rural
Tweets, n	407 million	350 million	18 million
Words per tweet, n	10.47	10.44	10.54
Unique hashtags, n	474,124	333,177	30,080
Hashtags per tweet, n	0.18	0.18	0.17
Word2vec model	all-tweets-w2v	urban-w2v	rural-w2v