NFL Tweet Dataset Version 1.0 August, 2013 This dataset includes .csv files containing tweets from the 2010-2012 NFL seasons annotated with relevant game information, partitioned into 1) weekly tweets and 2) postgame tweets. Weekly tweets are those that occurred at least 12 hours after the start of the previous game and 1 hour before the start of the upcoming game for their assigned team. Postgame tweets are those that occurred between 4 and 28 hours after the start of the previous game for their assigned team. A tweet was assigned to a team if the tweet contained hashtags corresponding to exactly one team. The file teams.nfl.hashtags contains a list of hashtags and corresponding teams. The game data was obtained from NFLdata.com and includes the home and away teams, their points scored in the game, the point spread line, and the total points line. The list of fields included in the *postgame.csv and *.weekly.csv files are contained in the tweets.key.csv file. See Sinha et al. (2013) for more information on how the dataset was collected and constructured. For more information regarding the tweets used to prepare this dataset, please contact one of Shil Sinha (shil.sinha@gmail.com), Chris Dyer (cdyer@cs.cmu.edu) or Kevin Gimpel (kgimpel@ttic.edu). The data is released under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported (CC BY-NC-SA 3.0) license. The license is detailed here: http://creativecommons.org/licenses/by-nc-sa/3.0/ Please cite the publication below if you write any papers using this dataset. Shiladitya Sinha, Chris Dyer, Kevin Gimpel, and Noah A. Smith. "Predicting the NFL using Twitter," ECML/PKDD 2013 Workshop on Machine Learning and Data Mining for Sports Analytics, 2013. bibtex: @inproceedings{sinha-13, author = {S. Sinha and C. Dyer and K. Gimpel and Smith, N. A.}, title = {Predicting the {NFL} Using {Twitter}}, booktitle = {Proc. of ECML/PKDD Workshop on Machine Learning and Data Mining for Sports Analytics}, year = {2013} } The paper is also included with this download (sinha+dyer+gimpel+smith.mlsa13.pdf).