The data contains about 50 millon tweets per month.

The data is composed of the following blocks of 3 lines:
T       2009-06-11 16:59:45
U       http://twitter.com/ibbored
W       @amberback #squarespace does? Hot damn. Now I want to win more.

T ... time of the tweet
U ... user that posted the tweet
W ... text of the tweet (includes links and hashtags)

Note that Twitter users use hash-tags (#squarespace), reply to one
another using @ sign (@answerback) and also put hashed urls into the 
posts.

The Twitter data was prepared and extracted by Jaewon Yang.

If you use these data cite:
J. Yang, J. Leskovec. Patterns of Temporal Variation in Online Media.
In Proc. Forth ACM International Conference on Web Search and Data
Mining  (WSDM '11), 2011.