This data comes from a scrape of the Twitter social network conducted by the Monkeywrench Consultancy. The full scrape consists of 40 million users, 1.6 billion tweets, and more than 1 billion relationships between users.

This dataset is a list of stock tokens and keywords collected from tweets sent between March 2006 and March 2010. The convention of using the token $STOCKNAME or $$ in a tweet has arisen on Twitter to talk about stocks, companies, and other information relevent to the financial industry. This dataset collects all uses of these tokens and extracts keywords from the associated tweet. Each stock token is listed with a timestamp, a tweet id, and any keywords found in the text of the tweet. These items are listed for each stock token contained in the tweet.