A new catalyst of civil uprising many believe has been identified in the form of social media. Examples range from the election protests in #Iran, the ousting of #Mubarak, the #ArabSpring as a whole, #WikiLeaks with its #Cablegate and the latest showcase, the #Occupy events in the United States. The matter whether social media like Twitter and Facebook actually contribute to these forms of civil disobedience or are just a form of ‘clicktivism’ has been discussed to a great extend. More recently, the debate has turned against Twitter with claims that the social network had a hand in some hashtags not becoming a trending topic.
It goes without saying that this is the kind of stuff for conspiracy theorists: the government is aware of the power of social media and is quietly instructing Twitter to suppress certain hashtags from making it to the trending topics list. Then again, the way in which 700 protesters were arrested on the Brooklyn Bridge shows an unrelenting crackdown on civil disobedience by the American government. In addition, the track record of governments around the world, including the United States government, shows plenty of violation of digital privacy and intrusive online behaviour. So, you wouldn’t really put it past them either.
Recently the question has been whether Twitter censored #OccupyWallStreet. Using the statistic-based trending analytic tool ‘Trendistic’, PhD student Jonathan Albright plotted #OccupyWallStreet on October 1, at the peak of its activity to date, against two of the trending topics: #WhatYouKnowAboutMe and #October. The statistics shown below cover the exact same time frame when #WhatYouKnowAboutMe and #October go trending whilst #OccupyWallStreet does not.
Twitter has dealt with trending issues in a blog titled ‘To Trend or not to Trend’. The reasoning behind some topics trending and others that don’t is that one must distinguish between a hashtag’s ‘popularity’ and its ‘breaking’ properties. Twitter apparently favors novelty over popularity: topics break into the trends list when the volume of tweets about that topic at a given moment dramatically increases, whilst topics don’t become trending or cease to be trending when the velocity of conversation isn’t increasing quickly enough relative to the baseline level of conversation happening on that day. This is also explained in the article ‘Data reveals that occupying twitter trending topics is harder than it looks’ and is accompanied by a graphical representation.
A striking discrepancy exists between the two graphs shown above, resulting in the fact that the logic concerning the distinction between ‘popularity’ and ‘breaking’ does not seem to be applicable in the case of the graphs by Trendistic, which clearly show a dramatic increase of the topic #OccupyWallStreet. It however does seem to be applicable in the second graph provided by SocialFlow.
It should be noted that Twitter’s algorithms are not public, as if they were, they would be constantly ‘gamed’. Sean Garrett, Vice-President of Twitter Communications explains that SocialFlow pays to have access, as other third parties and researchers do, to all of Twitter’s public data which is called the “firehose”. According to Garrett Trendistic has access to what is called the “gardenhose”, which comprises 10% of tweets.
It could be said then, that this is the cause for the discrepancy in the graphs, and that the fact that #OccupyWallStreet is not trending is indeed caused by the difference between ‘popularity’ and ‘breaking’. The problem is that you have to take Sean Garrett’s word for it as long as Twitter’s algorithms remain undisclosed from the public.