Sometime in the last few days my Twitter follower count clicked over a round number of thousands. In parallel, DeWitt Clinton published Sampling Twitter, some nifty research into how many Twitterers are really active. Being, obviously, an egomaniac, I was more interested in how many of the people following me are “active”.
I decided that an interesting first-cut approximation to “active” would be “have posted during the last sixty days”. Based on that metric, assuming my dinky Ruby script is correct, then as of this Monday, of my 4,089 ostensible followers 3,501 are “active”.
Of course, inside Twitter, they could compute real truths; answers to questions like “On average, how many different Twitter timelines in any given week are downloaded that include a post from person X’s feed.” But this is better than nothing; quite indicative actually.
Comment feed for ongoing:
From: David Warde-Farley (Jan 12 2009, at 15:36)
I had a similar inkling about a week ago. Being a rather low-profile Twitter user I was always curious, when I had a new follower, how they'd found me -- if we had contacts in common, etc. Sometimes this is extremely non-obvious and Twitter's web interface provides nothing like Facebook's "friends in common" feature. So I whipped this up:
http://discriminant.org/twit.py
Nothing exceptionally fancy, but it gets the job done (with 4 API calls).
[link]
From: Henrik Lied (Jan 13 2009, at 05:47)
Would you mind terribly to release the script? It'd be interesting to see how many of my followers are active.
[link]
From: Norman Walsh (Jan 26 2009, at 05:13)
What you really want to know, I suspect, is how many folks are actually reading your tweets. That's not necessarily correlated with tweeting.
I've been phenomenally busy the past few months. Though I've probably tweeted several times a week, there have definitely been weeks when I've left the reader off more than on.
If you miss more than 200 tweets, you can go back and find the ones you missed, but I rarely do. Where rarely is roughly equal to never.
[link]