Overview

As a first step towards designing effective mechanisms to thwart spam in Twitter, we study and characterize the linking strategies adopted by spammers in the Twitter social network.

Methodology

We identified more than 40,000 spam-accounts in Twitter and extensively analyzed the social links of these accounts. Our key insight is that spammers are able to acquire large number of followers in Twitter, primarily because of the presence of a set of users who engage in link-farming in order to acquire social capital.

Paper

1. •   Understanding and Combating Link Farming in the Twitter Social Network
   

      Saptarshi Ghosh, Bimal Viswanath, Farshad Kooti, Naveen Kumar Sharma, Korlam Gautam,

      Fabricio Benevenuto, Niloy Ganguly and Krishna Gummadi

      Proc. ACM International Conference on the World Wide Web (WWW), April 2012.

      [PDF] [BibTex]

Dataset

An anonymized topology of the Twitter social network is already available, for details see this page.

The dataset we used in our study includes 41,352 spammer accounts which:

1. •   were suspended by Twitter, as of February 2011, and
   

2. •  had posted in the life-time at least one shortened URL which was blacklisted by the bit.ly or tinyurl services.
   

If you are interested in the spammer data, please send us email to twitter-contact@mpi-sws.org.

Members

  Faculty

1. •   Krishna Gummadi
   

2. •   Niloy Ganguly
   

3. •   Fabricio Benevenuto
   

   Students

1. •   Saptarshi Ghosh
   

2. •   Bimal Viswanath
   

3. •   Farshad Kooti
   

4. •   Naveen Kumar Sharma
   

5. •   Gautam Korlam