Tweets Have Become Shorter Since 2009, Say Computer Scientists
Back in a aged days, contend in 2009, Twitter was a comparatively different amicable network that was beginnign to widespread like wildfire. In 2007, Twitter users posted some 400,000 tweets per quarter, by Jun 2010, they were posting 65 million any day. Today, there are 200 million purebred users who send around 400 million tweets any day.
During this brief time, Twitter has turn so renouned that a technical cant has entered a common language. Words like hashtag and @name would have seemed small some-more than nonsense usually a few years ago. But even a word twitter is now a noun strictly recognized in a Oxford English Dictionary.
Now justification is rising that Twitter competence be carrying a some-more surpassing impact on a approach we communicate. Today, Christian Alis and May Lim during a University of a Philippines contend they have totalled how a length of tweets have altered between Sep 2009 and Dec 2012 and say that tweets have shrunk dramatically in that time. “People are communicating with fewer and shorter words,” they say, roughly positively since we’re all regulating lingo some-more effectively.
They’ve also complicated how a length of tweets varies geographically in a US from state to state. And they’ve looked to see how a changes relate with several socio-economic statistics from a US Census Bureau. The formula are surprising.
First, some-more about a investigate itself. Alis and Lim collected 229 million tweets published between 18 Sep 2009 and 14 Dec 2012. They afterwards counted a length of any twitter and plotted this opposite a date of publication.
The information is interesting. The placement of twitter lengths has dual peaks. One is nearby a 140 impression extent for tweets, that Allis and Lam appreciate as a forced constraint. In other words, tweeters can’t send messages longer than this, even if they wish to, and so are forced to finish their messages during this length.
The second rise is what varies over time. Between Nov 2009 and Dec 2012, Alis and Lim contend a median tongue length in difference decreased from 8 difference to 5 words
This cutting is a tellurian materialisation that is loyal for a dataset as a whole, for tweets in English alone and even when all a links are private from a dataset.
(In Oct 2011, Twitter introduced a link-shortening algorithm that translates all URLs into 20 impression addresses. This causes a spike in difference that are 20 characters in length though doesn’t change a trend towards shorter tweets.)
Having found this extraordinary cutting effect. Alis and Lim ask what competence be causing it. “The shortening, it seems, can be explained by increasing use of jargon,” they say.
That’s poignant since it implies that Twitter users are apropos segregated into well-defined groups of those who know a same jargon.
In a extraordinary twist, Alis and Lim also investigate a subset of 800,000 tweets that are geotagged with a US state. They contend a numbers of tweets from any state are strongly correlated with a race of that state as available in a 2010 census.
In plotting a length of utternaces on a map of a US, a transparent geographical trend becomes obvious. “Southeastern and eastern US states tend to have shorter tongue lengths,” they say. Just because this is a box isn’t clear.
Alis and Lim go on to check for correlations with 51 variables totalled in a 2010 census and now published online. These are factors like “Persons 25 years and over who are high propagandize graduates or aloft from 2007 to 2011 in percent” or “Owner-occupied housing units in percent of sum assigned housing units from 2007 to 2011” or “2011 proprietor Black race in percent”.
It turns out that a usually non-static that correlates strongly with shorter utterances is Black race percentage. Why this should be isn’t transparent though Alis and Lim indicate out that there is justification that a black race uses Twitter significantly some-more than other groups and that lingo competence be some-more common in this group.
Of march association does not indicate causality. One approach to investigate this in some-more fact would be to inspect a calm of a tweets in detail. But Alis and Lim contend this is over a range of their study.
What’s engaging about this work is that conversations simulate a existent norms of a language. So any justification of change is significant. It competence be that we’re all regulating most some-more lingo in a tweets and that this represents a elemental change in a approach we communicate. Then again, we competence usually be training how to use Twitter some-more efficiently. At some level, maybe those things are even equivalent.
Your views greatfully in a comments section.
Ref: arxiv.org/abs/1310.2479: Spatio-Temporal Variation of Conversational Utterances On Twitter