It’s nice to be listened to. I guess. Maybe. Though I now find myself wondering whether I wouldn’t be better off shutting up.
I saw the first reports of Michael Jackson’s death on Twitter around 6pm. I ran a little script I threw together some weeks ago called “twitcent” to see just how many tweets would share the news. Twitcent takes advantage of the fact that Twitter gives a unique, sequential ID to each tweet to estimate the intensity of posting around certain terms. It retrieves a page of 100 search results for a particular search term – say “Michael Jackson” – and looks at the ID numbers of the first and last tweets listed. Take the difference of those numbers, and you get how many tweets were posted between search result #1 and #100. Divide, and you’ve got a percentage of tweets on the system in a discrete, small interval mentioning the term.
Is it accurate? I dunno. If my assumptions are right, it should be – if Twitter’s not always numbering sequentially, or if some large percent of tweets on the system are unsearchable, less so. Anyway, I ran several search terms through the engine and saw something I’d never seen before – search terms registering in double digit percentages, and the term “Michael Jackson” appearing in 13 – 20% of the tweets.
So I tweeted the following: “My twitter search script sees roughly 15% of all posts on Twitter mentioning Michael Jackson. Never saw Iran or swine flu reach over 5%” And then I went to make dinner.
Geez, think these guys read each other much? I’m flattered, I think. But worried that I’m now going to be quoted for the next several days as an “expert” on Michael Jackson twittering, especially as the NYTimes piece identifies me as a Berkman Center researcher.
Of course, by the time I’d gotten back online, the initial fervor had died down – here’s what my script turns up now:
2.152 % Michael Jackson
2.634 % jackson
2.242 % michael
0.312 % micheal
1.596 % MJ
0.119 % #MichaelJackson
That’s a lot of tweets, but now in the neighborhood of a busy swineflu day or the heart of Twitter’s interest in the Iran protests. What was interesting to me was the way the information flashed across Twitter, briefly bringing on the failwhale for some users – with one in seven or so tweets mentioning the death, it’s interesting to wonder whether people saw themselves as spreading the news, or as simply expressing shock, surprise, or their personal reaction. (And yes, I tweeted an update that the term was now down to roughly 3%. That one hasn’t gotten retweeted…)
What’s really interesting to me is the extent to which news reporters seem to have chosen Twitter as the go-to source for reactions to news events. It makes sense – there’s a premium in the news business on speed, on having a story faster than anyone else does, so the need for the quick quote makes Google hours to slow to help you. And the 140 character limit guarantees that whoever you quote will be pithy and limited to a single soundbite.
This, in turn, also increases the chance that you’ll be wrong. A proper quote from me would probably have been something like: “The search string ‘Michael Jackson’ is getting intense interest on Twitter at the moment, showing up in between 13-20% of tweets. It’s unlikely this level of intensity will continue through the night, but at the moment, it exceeds the intensity I’ve seen on Twitter during slower-breaking stories like #swineflu, #pman and #IranElection.” That, unfortunately, is 337 characters – far too long for anyone to read anymore. And a clarification in the form of a blogpost? That’s so 2006.