As a Canadian overseas, I can’t say that I want to perpetuate news about Toronto’s Mayor Rob Ford, since he is often one of the main topics that people bring up in relation to Canada. However, as he’s still making headlines and causing a stir on Twitter, I thought a Ford story would be a good way to share a slice of my latest learning about ‘big data’ methods and analysis.
With the purpose of trying out some new tools and ideas, I collected tweets about Toronto’s WorldPride festival, which took place this past June. It was a huge shindig and while I wasn’t able to capture every relevant tweet, the 6 hashtags that I tracked* (#WP14TO, #WorldPride, #PrideToronto, #TorontoPride, #PrideTO, #WPTO14) turned up a pretty good dataset totalling 68,231 tweets. This dataset showed some cool trends relating to participation, especially people’s awesome selfies and photo documentation of the WorldPride parade (check out the National Post’s photos if you’re lacking rainbows in your life). I hope to eventually share some of these broader analyses but today I just wanted to look at a little bump that showed up after the festival, circled in Figure 1.
Figure 1. Total WorldPride tweets over time
This little spike of nearly 1000 tweets happened when Toronto’s Mayor, Rob Ford – fresh out of rehab, as all the latest news stories note – refused to join in the standing ovation at a city council meeting to thank WorldPride’s coordinators. That’s right, everyone else stood up and clapped but Ford, with his history of avoiding Toronto Pride and opposing visible support for LGBTQ people throughout the city, remained seated. Apparently, to add insult to injury, all of this came alongside Ford casting the only vote against launching a study to determine if more homeless shelter space for LGBTQ youth is needed in Toronto.
So what did Torontonians do? Well, when the incident first happened, some of the city councillors tweeted about it. This is reflected in the first bump in Figure 2, when many people retweeted these preliminary expressions of disappointment with Ford’s behaviour. Figure 2 shows the volume of tweets over time for the bump that was circled in Figure 1 but here I’ve also plugged a bit of code into Tableau to show the different types of tweets. You can see that this whole Twitter event was characterized by people retweeting, often using the popular #TOpoli (Toronto politics) alongside the WorldPride hashtags.
Figure 2. Rob Ford incident over time, sorted by tweet type
The mainstream press caught wind of the story and a bit later in the day, CBC News tweeted about it, adding a photo of Rob Ford sitting during the applause. However, the real kicker in terms of momentum happened when media personality Jian Ghomeshi (broadcaster, musician, host of Q) made a tweet that resonated with a bunch of people:
— jian ghomeshi (@jianghomeshi) July 9, 2014
Okay, so Ghomeshi’s tweet wasn’t an original, he simply added his own opinion to the CBC’s previous tweet. But the combination of celebrity critique with the compelling visual made this the most popular retweet of the whole debacle, raking in nearly 300 retweets in my dataset and gaining even a few more that weren’t captured during my data collection.
What does it mean that retweets dominated the dialogue throughout this whole spectacle? Does it show that mainstream media still has the loudest voice even on social media platforms, which are often lauded as being participatory and democratizing? Perhaps. Does it mean that Torontonians are lazy and would rather just press the ‘retweet’ button than weigh in with their own opinions? I think not.
Retweeting IS a form of participation (boyd, Golder & Lotan, 2010). It serves multiple purposes: it gets the word out by making a conversation more visible, it engages a wider network of participants in the dialogue, and it shows support for a particular viewpoint. Ghomeshi’s tweet hits the important points – it expresses a negative sentiment for Ford’s actions and drives it home with visual evidence of his non-participation. People who retweeted likely felt that this tweet represented their feelings accurately. It’s also likely that a broader range of people feel comfortable retweeting something fairly political when it’s led by a media personality because they may not be ready to make such strong statements independently.
A couple of the participants in my MSc research who weren’t out to their families talked about this. They explained that they wanted to show support for LGBTQ people and did so through political tweets that didn’t reflect their identity as much as personal statements. It seems that retweeting might be a way for a lot of people to get involved and stand in solidarity with a certain viewpoint without their actions implicating them beyond their capacity. Our personal situations may not always allow all of us to be highly vocal activists, but retweeting could add power to those who do speak up so that they speak on behalf of a collective – a collective of Twitter users, at least.
Personally, I might also guess that users mostly retweeted during this incident because, well, is there really anything left to say about Rob Ford?
- I’ve added Tableau to the “Assorted tools” page in case you’d like to have a closer look at it. Their website allows a free trial along with some great video tutorials.
- A good resource for what/why/how to work with Twitter data is the book “Twitter and Society” edited by Katrin Weller, Axel Bruns, Jean Burgess, Merja Mahrt and Cornelius Puschmann.
- You may have noticed that we’ve been talking about ‘big data’ without heaps of numbers and statistics. While this speaks to my tendency toward qualitative research, it’s also a technique from the digital humanities methods that I’ve been learning about. It’s possible to take large sets of data and do a ‘distant reading’ (Moretti, 2007) of them in their entirety (like Figure 1) and then to drill down into more qualitative types of content analysis. I turned to Richard Rogers’ book “Digital Methods” as inspiration for this.
- Disclaimer: This was just an exercise (with a relatively small number of tweets!) that I’ve presented for discussion – there are of course lots of limitations to ‘big data’ analysis and the use of Twitter data. While I don’t address these here, other people have – start with boyd and Crawford’s “Critical Questions for Big Data” to get a handle on the issues.
*All of this was done with the gracious help of QUT’s Social Media Research Group, especially with Jean Burgess’ ninja Twitter data collection skills and Darryl Woodford’s crash course on Tableau analysis for Twitter data.
In text references:
boyd, d., Golder, S., & Lotan, G. (2010). Tweet, tweet, retweet: Conversational aspects of retweeting on Twitter. Proceedings of the 43rd Hawaii International Conference on System Sciences, IEEE. doi:10.1109/HICSS.2010.412
Moretti, F. (2007). Graphs, maps, trees: Abstract models for a literary history. London: Verso.