The Twitter Project Page at MPI-SWS

Welcome to our Twitter project page. The data used in our ICWSM'2010 paper is available for use by the wider community.   Based on Twitter's explicit request, we are only sharing the anonymized topology of the Twitter social network. Please understand that we are not allowed to share any tweet information.   If you have any questions please send us an email at    twitter-contact (at) mpi-sws.org.

Papers

• Meeyoung Cha, Fabricio Benevenuto, Hamed Haddadi, Krishna P. Gummadi,
  The World of Connections and Information Flow in Twitter,
  In IEEE Transactions on Systems, Man and Cybernetics - Part A. Volume 42, Number 4. July 2012. Pages 991-998.
  [Download Paper (523KB)]

• Jisun An, Meeyoung Cha, Krishna P. Gummadi, Jon Crocroft,
  Media landscape in Twitter: A World of New Conventions and Political Diversity,
  Proc. International AAAI Conference on Weblogs and Social Media (ICWSM), July 2011
  [Download Paper (436KB)]

• Meeyoung Cha, Hamed Haddadi, Fabricio Benevenuto, Krishna P. Gummadi,
  Measuring User Influence in Twitter: The Million Follower Fallacy,
  Proc. International AAAI Conference on Weblogs and Social Media (ICWSM), May 2010
  [Download Paper (205KB)]

Data Characteristics

Data from our ICWSM 2010 paper is available from the link below.  Our datasets have been anonymized to protect the privacy of the users themselves.  We are only releasing information about the Twitter link structure.

Note that we are unable to release any non-anonymized data.

Data set

• List of links:

  This file contains a list of all user-to-user links that we crawled from Twitter, based on the snapshot of the Twitter network in September, 2009.  The file contains 1,963,263,821 directed social links.

  Format:  Gzipped ASCII.  Each line contains two user identifiers, implying a link was observed from the first to the second users (first user follows second user).

  Data: Twitter follow links (10.73GB)

• Time series of retweeting conventions:

  This file contains number of new adopters of 7 different retweeting variations (RT, via, Retweeting, Retweet, HT, R/T, and the recycling symbol) per day.

  Format:  xlsx.   Each line corresponds to a retweeting variation, where the first column gives the name of the variation and the subsequent columns give the number of new adopters per d

  Data: Retweeting conventions time series (< 1MB)

We are also sharing a set of spammers nodes within this graph from a related project about link farming in Twitter.

Media Coverage

• ReadWriteWeb blog article by Sarah Perez on March 19th, 2010 (link).
  The same article was picked up by the New York Times (link).

• Interview article by Brent Lang at theWrap on April 12th, 2010 (link)

• A research blog article at Harvard Business Review by editor Scott Berinato on May 7th, 2010 (link)

Applications

• Media Explorer - Visualize media landscape in Twitter
• Whom to Follow - Discover topic authorities on Twitter
• Who is who - Discover crowdsourced opinion on Twitter users
• What is happening - Discover topical content on Twitter
• Information Trade - Characterizing Information Trade on Twitter

Members

If you are using our dataset, please use the following BibTeX entry to cite our work.

@inproceedings{icwsm10cha,
  author = {Meeyoung Cha and Hamed Haddadi and Fabricio Benevenuto and Krishna P. Gummadi},
  title = {{Measuring User Influence in Twitter: The Million Follower Fallacy}},
  booktitle = {In Proceedings of the 4th International AAAI Conference on Weblogs and Social Media (ICWSM)}},
  month = {May},
  year = {2010},
  address = {Washington DC, USA}
}

We are also sharing a set of spammers nodes within this graph from a related project about link farming in Twitter.