How would you pitch ‘Did Twitter censor Occupy Wall Street?’?
Did Twitter Censor Occupy Wall Street? was written and headlined intentionally provocatively as it was intended to break the cycle of “trending” conspiracies by going directly after the conspiracy itself with basic statistical data and a hard-angle investigative journalistic approach.
The article questioned the perceived role and transparency of social media, voicing concern over trust of commercial algorithms such as Twitter’s trending topics, and pointed to the specific instance of #occupywallstreet censorship. It then focused on easily interpretable and hard empirical evidence.
Using graphs and first-person narrative, Did Twitter Censor Occupy Wall Street contradicted popular myths around the #occupywallstreet, #occupywallst and #ows censorship claims, helping to generate awareness on how topics actually “trend” on Twitter. This in turn helped to further stir the debate around social media’s role in activism in general.
Although “Did Twitter Censor Occupy Wall Street?” was written with the intent to question the perceived transparency of social media in general, its popularity resulted in something unexpected: It marked the first time Twitter directly responded to the censorship debate. Soon after the piece went live, Sean Garrett, Twitter’s SVP of Communication, began replying directly in the comments to readers.
Sean disclosed that at the time no one, save for the US Library of Congress, had access to Twitter’s “fire hose” or full data feed, clarifying for the first time that third party analytics websites typically only have access to around 5%, or the “garden hose”, a small fraction of the full Twitter data set. Sean also defended Twitter’s role and reiterated non-censorship.
The data feed access clarification was a breakthrough, since Sean commented, there was no direct public reply from Twitter regarding third parties’ access to just the “garden hose”, which utilise less than 10% of Twitter’s data and extrapolate. This disclosure alone, as well as the personal attention from a Twitter employee, was a powerful tool in helping to generate real information around the Twitter conspiracy theories.
What inspired you to make ‘Did Twitter censor Occupy Wall Street?’?
Activists online frequently assume that social media, especially mainstream services such as Twitter and Facebook, are natural advocates of their cause because of social media’s unique capability to quickly ‘confirm’ or display efforts around activist coordination efforts. With regards to “trending” hashtags as well as leading keyword search result placement, this is not always the case.
At the time when “Did Twitter Censor Occupy Wall Street?” was published, no direct journalistic challenges to Twitter’s trending censorship claims existed from credible media sources using hard data.
Popular search results and news links at the time included: 1) a somewhat conspiracy-oriented and buggy ‘trending’ geographical map which claimed ‘censorship’ based on the absence of visual ‘trending’ keywords; 2) links to previous articles from unrelated events past ‘debunking’ trending topics on Twitter and 3) links to more involved data analytics projects which either did not address #occupywallstreet directly or were difficult to understand and interpret for mainstream news audiences.
Second and more importantly, there was no official response by Twitter to #occupy and #occupywallstreet trending censorship claims, which were gaining conversational momentum quickly. There was only a great deal of perpetuating misinformation, nonsensical hijacked comment threads, and battles of opinions and fact in the hashtag feed.
Did you work by yourself or in a team?
This project was undertaken individualy, with the copy editing assistance of Meg Clement, the Politics and Society Editor at theconversation.edu.
How did you get a hold on the data you needed?
At the time, there were countless Twitter ‘trending’ websites available, which were of little use as most have no data to back up their ‘rankings’. In addition, there were various professional online services for social media monitoring . These subscription-based services are expensive to access and typically focused toward public relations.
Journalists, especially those without a budget, do not typically have access to nor the skill set needed to utilize these on demand. This left a single tool available for freely open-source use: Trendistic. Using Trendistic, the only freely available Twitter analytics tool at the time with statistics and percentages, #occupywallstreet, by far the most influential Occupy Wall Street Twitter hashtag, was plotted on October 1 (its peak activity at the time) against two other keywords – #whatyouknowaboutme and #october – which trended exactly at the same time.
Each graph was then exported using a screenshot plug-in and combined in Photoshop into a composite image. The graphs then represented the statistics over the peak of the #occupywallstreet censorship debate on Twitter and Facebook, when #occupywallstreet appeared to go viral on Twitter (directly before and during the NYPD Brooklyn Bridge mass arrests and protest)
Which tools were used making this production?
Trendistic and Photoshop. Trendistic, as a Twitter analytics tool was used for three reasons: 1) Trendistic was the only available Twitter analytics service at the time that contained hard statistical data; 2) It was completely open and free to use; 3) Trendistic is highly credible as the parent company provides the search functionality and indexing tools for Reddit.com, one of the leading social network news sites in the world. Oddly, Trendistic is now offline.
How did it take to complete the article?
Approximately three to four days.
Were there any bumps in the road?
The author is a busy graduate student in media with little to no support in terms of budget or resources. Further, the site this piece was published in Australia’s The Conversation, a brand new experiment in academic journalism. At the time, the site was just launced, so the reach and SEO of the piece was unknown. This was inherently challenging. In addition, according to the Conversation’s guidelines, the article had to be “pitched” to the section editor before even being undertaken.
The production ended up coming together well as The Conversation is setup like a professional news organization with high standards: stict editorial oversight, copy editing, as well as a system for obtaining relevant Creative Commons photos for the piece.
Do you have a useful tip for starting data journalists?
1) Credible data and research tools are freely available if you look hard enough.
2) It doesn’t require a vast amount of statistical or mathematical knowledge to successfully challenge conspiracies or established ideas.
3) Just get the story out there and worry less about the medium. It will probably catch and spread if it’s relevant and well-written. You can create an entire awareness campaign from pieces less than one thousand words.
4) You never know who might end up engaging with the piece!