Starting off The Statbot, I analysed Robert Scoble’s twitter stream. Scoble never lets you down, especially when it comes to providing data: He’s on twitter, friendfeed, flickr, qik, a blog and god-knows-how-many-other-services. I will be analysing the rest of his “Online Life”, with posts about them on regular intervals, starting now with his Twitter Stream.
Size
The dataset used consists of 10,598 tweets from http://twitter.com/scobleizer, spanning 523 days (till April 27th 2008, when this was written), with 175,543 words, 1,026,899 chars, and 1,095,406 keystrokes! Means more than a million keypresses had been sacrificed by Scoble for twitter. These came at an average of 20 tweet per day (explained below), 16.5 words per tweet, 97 chars per tweet and 103 keystrokes per tweet. Still, just about half as many chars as found on his Wordpress blog when I last profiled it more than a year ago.
Tweeting Frequency
Here’s the chart showing the No. of Tweets per day:

For almost four months after starting the account on November 20th, it remained dormant: Until the end of February, there were only 114 tweets, at a dismal average of 1.1 tweet a day. There’s another slow period during May-Jul 2007, and another one between Mid Oct-Dec 07. In short, he’s totally erratic – see the huge spikes followed by the flat and low worms. He is going ballistic in 2008 though – He has 4693, or 44% of his tweets, in 2008, at an average of 40 per day, which is double his overall average
Clients Used

About half the tweets come from the web, and about a third come in via IM. Twitterific is the most used Desktop Client (Mac only. Ugh). Only 3 SMS Messages. He tried twhirl, iTweet and Snitter for a bit, but probably he doesn’t like them. My theory is that he likes IM because you are not limited to just 20 tweets per minute. Being restricted to 20 tweets per minute will make you lose a lot of tweets, especially when you follow 21k people. I find twhirl dropping many tweets, and I follow only 1.5k people. I wonder how he managed with the Web though (most of his recent tweets are via IM).
Replies
Of those 10,598 tweets, 6,831 or approximately 64% are replies to someone else.

Extremely chatty, and not too many original thoughts
But I fancy I’ll find someone chattier sooner than later
Those 6,831 replies are to 2,260 people, at an average of 3 replies per person. Here’s the top 10:
| Rank | User | Replies |
| 1 | spin |
150 |
| 2 | loiclemeur |
142 |
| 3 | davewiner |
127 |
| 4 | techcrunch |
99 |
| 5 | prokofy |
76 |
| 6 | chrispirillo |
59 |
| 7 | duncanriley |
51 |
| 8 | gapingvoid |
46 |
| 9 | jasoncalacanis |
46 |
| 10 | steverubel |
41 |
I’ve never heard of spin before, neither of prokofy. I’m not too twitter-literate
Also of note is that Dave Winer is No.3, and has blocked Scobleizer…
Also, the replies are pretty Long Tailish – 36% of the replies are to the top 100, while 64% are to the bottom 2160.

That’s good, but here’s another chart:

Only about 25% of his followers get at least one @ reply. Conversationalist? Or maybe those followers don’t @ to him? I don’t know, because there is no data available (yet) on the number of @scobleizers in the twittersphere.
Here’s another pretty interesting chart:

There’s a 55% chance that you won’t get more than one reply from him (if you get a reply at all), and a staggering 89% chance you won’t get more than 5 replies. Couple this with the fact that 64% of his tweets are replies. His attention is nowhere near “undivided”
Trivia
Here comes the fun part

Mondays are the most active, followed closely by Sundays. I can understand Sundays, but Mondays? Maybe because Twittering is now part of his Day Job?
Also, he seems to be twittering at all odd hours…

I had quite a bit of trouble with this one, due to Time Zone differences. In the end, I just converted everything to PST, since that’s where Scoble is most of the time anyway.
As you can see, tweeting gets going at around 7 PM and goes on till 12 AM, and slowly dies down at 1 AM. Usually sleeps from 2 AM to 7 AM, and Twittering picks up again after 9AM.
Also, this very colorful chart (along with all the other charts here), is made possible because Steve Clayton sent me a copy of the awesome Office ‘07 (along with a couple of signed cards and a Blue Monster Tee). Thanks!
The End
I’m working on similar stats posts for other folks, and here’s the shortlist: Jason Calacanis, Chris Brogan, Jim Long, C.C Chapman, Karoli, Guy Kawasaki, Justine(simply because I like her avatar;)), Dave Winer, Hugh MacLeod and anyone else you may nominate in the comments. Also, I’ve got some very cool stuff coming down the pipe(such as an analysis of the Techmeme Leaderboard, creation of a Flickr Leaderboard, TechCrunch analysis, an Apple Weblogs showdown, etc), so don’t forget to subscribe to the RSS feed if you found this interesting
Update: Preview
Here’s a very small preview of what’s coming up next:

Notice which worm is on top? (ducks, and runs)
15 responses so far ↓
1 Damon // May 1, 2008 at 8:21 am
Sweet stats! For people that might want to check out their own, http://tweetstats.com has several of these statistics displayed above.
For Scoble: http://tweetstats.com/graphs/Scobleizer
2 Susan Mernit // May 1, 2008 at 8:50 am
Yuri..this is great fun and very interesting–keep up the good work. Louis, make sure this boy has enough caffeine to execute all your clever ideas!
3 louisgray // May 1, 2008 at 8:52 am
Susan, be assured that Yuvi is the creative one. All these ideas are his ideas. He’s the one behind the ideas and tech. I’m just here for support and late-night e-mail discussions.
Great debut, Yuvi!
4 Eric Rice // May 1, 2008 at 10:44 am
Heh, I’m a statistic, yay!
@spin
5 Robert Scoble // May 1, 2008 at 11:08 am
Eric (aka “@spin”) I enjoy our late-night talks.
I only single respond to a lot of people because I’m watching so many people and most people are very noisy (talking about their lunch, etc). Once in a while I see someone talk about something I’m interested in, so I respond, then they go back to being noisy.
I’m using Twirl again this week, but am missing Tweets (getting more work done, though).
6 Robert Scoble // May 1, 2008 at 11:08 am
BTW: thanks Yuvi, I always love your stats. I learn a lot about myself that way.
7 Scobleizer — Tech geek blogger » Blog Archive Early adopter angst « // May 1, 2008 at 12:07 pm
[...] we’re talking about Twitter, Yuvi, the wonderkid in India, did an analysis of my usage of Twitter that’s pretty interesting. Possibly related posts: (automatically generated)Meet the [...]
8 Alex Hammer // May 1, 2008 at 1:22 pm
You are a talented individual, as your reputation already affirms.
9 Webmaster Blog | Robert Scobles Tastatur… // May 1, 2008 at 1:29 pm
[...] ich nicht sein. Gut eine Million Tastenanschl
10 Simone Brunozzi // May 1, 2008 at 6:49 pm
This analysis is a good starting point, but I bet there is something more to discover if you’re able to confront Robert’s stats with other influencing people like LoicLeMeur or JasonCalacanis, for example.
You may want to check out my Twitter Power list of interesting sites and tools: http://www.brunozzi.com/en/2008/05/01/twitter-power/
I added The StatBot in the link section, of course, you surely deserve some link love
11 Simone Brunozzi » Blog Archive » Twitter Power! // May 1, 2008 at 6:51 pm
[...] Check out this statistic analysis of Robert Scoble’s tweets. Share and Enjoy: These icons link to social bookmarking sites where readers can share and [...]
12 Prokofy Neva // May 1, 2008 at 8:12 pm
@davewiner and @techcrunch blocked me because I criticized them.
I blocked @jasoncalicanis in protest when he said he’d give a free Mac Pro to a randomly selected Twitterer who followed him.
13 Yuvi // May 2, 2008 at 12:48 am
@Simone: Thanks
I started with @scobleizer simply because I can use him as a baseline when comparing others. Will.Do.More soon 
14 Jeremiah Owyang // May 2, 2008 at 5:09 am
Great analysis, I’d be interested in seeing similar analysis on my own twitter stream.
15 Timo Heuer // May 2, 2008 at 6:05 am
Your stats are great as always. If you want to to this for German twitterers too, try @NicoleSimon or @theuer (which is me ;)).
Leave a Comment