Details Of #iranelection Twitter Data
Results of #iranelection vs. #WorldCup vs. Bieber comparison (data collected over 2 hours on June 18th, 2010)
:
- #WorldCup: 33,166 Tweets spread across 23,396 Users
- Bieber1: 10,730 Tweets spread across 7,101 Users
- #iranelection: 897 Tweets spread across 133 Users
Initially, it's clear that there were siginificantly more tweets about the #WorldCup and Bieber than #iranelection; granted, this data was collected around six days after the actual anniversary.
However, more interesting is the data about the users who were involved in these conversations. There was a fairly even distribution of Tweets to users for the World Cup search - there were only 1.41 Tweets per user. Similarly, the Bieber baseline was 1.51 Tweets per User. In stark contrast, the Iran election conversation seems to allude that it is a small but vocal population - 6.74 Tweets per User were present in this group. Regardless of the lag time in actual data collection, this difference seems to underlie a major problem with making an argument for widespread conversations about the Iran election, one year later. While there are people talking about it, and talking very vocally; they have appeared to lose most of the steam of a wide base of appeal in the ever-changing Twittersphere, where a hot thing may only last a matter of hours.
Supporting the idea that this is a small yet vocal group is the number of "conversational tweets" (Tweets that are either a direct mention or a re-tweet (where the "in_reply_to_status_id" is an actual Twitter internal id) present in each dataset. Whereas the #WorldCup data yielded 1180 conversational tweets (3.58% of the total Tweets) and Bieber 990 (9.22%), the #iranelection data showed a strikingly less conversational situation - only 8 of the 897, or 0.89% of the total tweets, were directed at other users.
Additionally, from a previous data set collected in 2009 during the actual election, we can see that the same general demographics are likely at play because the time zone breakdowns (see graph) and account creation dates (see graph) look strikingly similar. From June 16th to October 24th, 766,263 Tweets containing the term "#iranelection" were collected for a previous study and showed the time zone breakdowns and account creation dates for a statistically representational portion of the #iranelection data.2
In other words, we are likely seeing the same users talking about the situation - furthermore, the volume for the World Cup and about Justin Bieber, as expected, was at a much higher volume than the #iranelection conversation. The Twitition's claims, then, would seem to be misinformed.
