Frequently Asked Questions


Q. The Word Offsets in Task A are incorrect. For example:

257343699460173824      10115042        23      25      positive        One ticket left for the @49ers game tomorrow! Don't miss the rematch of the NFC Championship game against the NY Giants! Hit me up!

The tweet only has 23 characters.

We are aware of the issue with the word offsets being incorrect. We are working on resolving this and will send out an update as soon as it is available.


Q. Can I get the test data from 2013?

We will be releasing the 2013 test dataset once we resolve the word offsets. We will be releasing it in the same manner as the training data.


Q. When I download the training data 1/3 of the tweets are no longer available. Can you email me the full dataset?

Some of the data is no longer available because the user has deleted their account content. Unfortunately there is nothing we can do about this. We cannot email you the full dataset. Sharing data is a violation of Twitter's terms of service. This is why you have to download the data on your own.

Contact Info


  • Sara Rosenthal, Columbia University
  • Alan Ritter, The Ohio State University
  • Veselin Stoyanov, Facebook
  • Preslav Nakov, Qatar Computing Research Institute

email :

Other Info


  • Join the Google group:
  • Download the data
  • See the FAQ
  • Results have been released here
  • The 2014 gold labels + scorer are here
  • The 2014 test dataset is here (combines five testsets: Twitter-2013, SMS-2013, Twitter-2014, Twitter-sarcasm-2014 and LiveJournal-2014)