5v3n.com

How to measure the weight of a twitter profile


You don’t need to be a scientist to realize that twitter is not to be categorized as a social network in the classic sense - but something else.

Twitter is not bi-directional: follower-followed vs. friend-friend). It mainly consists of links to sources of knowledge & nonsense. The rest are short thoughts and what people sometimes call the tourette’s syndrome of the web.

I love it ;-)!

But what you do need a scientist for is to realize what twitter could really be categorized as: a news network.

But wait, there’s more. I had some nice debates with @Kuemmel_HH about means of measuring the “weight” of a twitter profile:

  • Number of followers? No, there’s so many spam accounts that catch followers like flies - 20k followers, 20k followed, 0 weight.

  • The proportion of followers to the number of people the account follows? Nice, but there’s some tricky spammers out there who not only catch followers like flies but unfollow the followed after a while - 20k followers, 200 followed, 0 weight

  • The activity & impact? Sounds great! But what are the parameters???

@haewoon has some interesting answers to this question.

He analyses a complementary cumulative density function to measure the “simple weight” of a profile (pages 28 and up), plus explaining the nature of retweets based on a nice graph (pages 45 an following).

Got it? Impressive! If you’re not into math, just have a look at the resulting diagrams, they’re quite self explanatory ;-).

To put it short: there are some profiles he calls “super hubs” that have many followers & massive impact on the twitter-verse. Which is no surprise, but now it’s kind of official and what’s more important: you can put a figure on it.

Check out his summary & don’t miss the empirical RT trees on page 54 - it’s a pity that we can’t see them in a decent resolution:

Thanks to the guys at massklusive for pointing at the post at faz.net. It’s both german, so for the english talking audience: Guy Kawasaki is spreading thoughts about that study, too.

UPDATE: There’s another fine paper on the subject. Thanks to faz.net for drawing my attention to it. It’s also featured by the NY Times. Enough said, enjoy:
Meeyoung Cha, Hamed Haddadi, Fabricio Benevenuto, and Krishna Gummadi In Proc. of International AAAI Conference on Weblogs and Social Media (ICWSM), May 2010