Twitter Mania

Over the past few days, there have been SO many articles about Twitter in the mainstream press. It's funny, because it comes on the heels of our in class discussion on "The Tweet Life." This article addresses how the media, and the Washington/NYC corridor may be so much more obsessed with Twitter than other places.

It also makes another valid point: That it's really easy to use social media to explain phenomena that we would have explained some other way (I found my birth mom on Twitter, or Facebook...but people were doing that already).

Elizabeth pointed an article my way (using our #SMPASOCIAL hashtag!) about how Twitter can be used to map emotional moods. This is a study that's been picked up by the Washington Post, and the New York Times.

The findings are pretty interesting, but sort of what you would expect: that we get happy when we get off from work; but the underlying thing I find so fascinating is that this is a mapping of 84 different countries using Twitter from 2008-2010.

Another article - perhaps coincidentally, or perhaps on the tails of this study in Science, appeared as the large weekend section in the WSJ.

Things you can learn from Twitter data:

*perhaps mapping stock trends - and even hedge funds are getting into this (WSJ)

*mapping cross-cultural moods

*charting potential flu outbreaks

*tracking political activity and voting behavior

BUT - one must take social science for what it is. In the Science study, as Gawker (yes that's my citation for methodology) look at what the scientists did:

Cornell University sociologists used text software to detect the presence of positive words and icons (ie, “awesome,” “agree,” " :) ") and negative ones (“annoy,” “afraid,” " :( ") in 509 million tweets from 2.4 million users in 84 different countries over a two-year period. Researchers were very psyched to study people's emoticons

A few problems with that, as admitted in the NYT article.

How often do you say "I just woke up with bedhead. Awesome :)" on Twitter.

OR "The economy is doing super today. Great-O"

OR "Unemployment- my bag of tea ":)

These are, basically, variations of tweets I've seen.

So you really need to be careful with machine-coded content analysis, as it doesn't pick up sarcasm. One can guess though, that perhaps enough people aren't sarcastic so as to keep the study valid. But the researchers don't know.

Additionally, are people MORE serious about some things than other things (e.g. being interested in a particular stock vs. being interested an particular politician). Another potential question.

And another reminder: Twitter is really easy to use for scientific data. So instead of asking Verizon, say, for access to SMS (though as Morozov tells us, it's all out there) - it's very simple to get Twitter data through the public API.

So it's easy to monitor, and as we've talked about in the Tweet life article, what's your sample? The most media-addicted people in 84 countries.

I'm not convinced.