Zack: I total agree with you; there are a lot of spam accounts on Twitter, many of which have feminine names. However, I think that the researcher(s) chose to keep spam accounts in the data set because it holistically reflects the Tweet landscape since many of those accounts are amongst the most active contributors.

Steven: Yes, the lifestyle hypothesis is a generalization of the female gender, but so are the nurturing tendencies (Gossip Guys anyone?). After all, these are only fitting stereotypes and maybe the truth is that there is no significant difference between male and female behaviours within social networks…

You’re right, Twitter’s distribution is comparable to that of other networks but it’s surprising that 90% of Wikipedia’s content is even generated by the top 15% of contributors. To me, Twitter’s microblogging 140 character format poses less barriers for entry than editing knowledge based encyclopaedia articles.

I am also on the lookout for better social media analytics, I will share the link with you if/when I find something interesting

Giulio: Very interesting read! I especially agree with “public beats private”, “Simplicity beats complexity”, and “neighbourhoods beat networks”.

Conversely, I disagree with “circuits beat channels”. I find that tweets are still prevalently broadcasts. Oprah isn’t really conversing with anyone, most of the time she’s updating us on her life or promoting her show. Here’s a recent one : “Just saw RUINED…. OMG….. EVERY SINGLE PERSON ON STAGE WAS OUTSTANDING…. Boo-hooing next to Jane Fonda had to borrow tissue.9:18 PM Jun 5th from TwitterBerry”

“Neighborhoods beat networks. Twitter’s network effects don’t feel much like standard ones. I can subscribe to your feed, yet you don’t have to subscribe to mine — times millions. What’s going on here? Twitter realizes neighborhood effects, not just network effects: complex sets of intersecting, overlapping, mutually reinforcing network effects. Oprah’s followers are a neighborhood, and so are Ashton’s. You can benefit from joining many of these neighborhoods — not just one larger network.”

The full post is at: http://blogs.harvardbusiness.org/haque/2009/06/twitter_2.html

I am also not surprised about the distribution of tweets across members. It follows the usual pattern of networks, where 80% ( in this case 90%) of the value is created by 20% (in this case 10%) of the nodes. It would be interesting to see a more sophisticated analysis of tweet patterns.