> By Guillaume Valadon and Korlam Gautam
Twitter is a micro-blogging platform that allows its users to post short messages (aka Tweets) to be displayed on their profile pages. A social network is also embedded in Twitter: users can receive messages from users they follow. In the Twitter terminology, there are following and followers. Given a user A, his followers are the users who receive his messages, and his followings the users he receives messages from (similar to RSS feeds subscription). Note that the following/followers relation is not symmetric: if user A decides to follow user B, B does not automatically follow A back.
The data set used to make this plot was collected in June 2009 using our own Twitter exploration tools as well as the streaming API. It contains information about 1 427 116 Twitter accounts. The plot represents the correlation between the number of following and the number of followers. For a given following/follower value, the color describes the number of accounts that have this value. For example, we collected information about 10 accounts with 10 followings and 10 followers.
An interesting result can be seen for small followings and followers values: 67 % of users have followings and followers values strictly below 100. Among these users, we can clearly see the common likelihood to have the same number of followings and followers.
Finally, we can clearly identify the well-known following/followers ratio that forbids a user to have more than 2000 followings if it has less than 2000 followers (that’s the horizontal part of the red line). From this preliminary investigation, it seems that after 2000 followings, the followings/followers ratio linearly grows with the number of followings (that’s the growing part of the red line); this fact was not known before the study, and it is still not advertised by Twitter.