This is the fourth in a series of five posts in which I am sharing analysis of how the the recent federal election (#elxn42) played out online. While I expect I will write a fair bit about the election over the coming weeks and months, this series specifically shares data analysis covering the entire election period.

In this post I look at demographics of the online participants as well as conduct an examination (and comparative analysis) of the Twitter followings of outgoing Prime Minister Stephen Harper and Prime Minister elect Justin Trudeau. I’ve excluded leaders Thomas Mulcair, Elizabeth May and Gilles Duceppe only because of the time necessary to conduct a meaningful analysis. I decided on the comparison rather than a strict election analysis since this approach can lead to some interesting discussions.

Analysis was conducted using Sysomos MAP and This analysis only considers mentions based on keywords; it does not consider sentiment.

Be sure to follow the digital activities of our newly-elected MPs on Our Digital Parliament.


  • Students dominated growth in Twitter followings
  • The online discussion was mobile
  • Twitter audiences listen more, contribute less
  • Participation skewed male
  • Ontario led the country, Canada led the world

Students dominated growth in Twitter followings is an interesting tool and, in this post, I’ve only scratched the surface on the breadth and depth of analysis I can conduct with this analysis. Note that I intend to do more with in the coming months as I conduct ongoing analysis of the role of digital in the 42nd Parliament. automatically identifies segments of Twitter followers based on affinities. The affinities can be all over the place inasmuch as the segments run the gamut of interests (e.g. sports, music), stage of life (e.g. students, moms), geography (e.g. India, Quebec, Calgary — yes, countries, provinces AND cities), and more. The groupings are based on densities of data and metadata.It’s the job of the analyst to examine the segments and then identify/name them. There is often a “miscellaneous” segment made up of people who have share no dominant characteristics. This is often the last and largest segment within the specified limit of segments.

The following pie charts compare the identified segments associated with Stephen Harper and Justin Trudeau on October 22. It’s interesting to note that Mr. Harper’s largest identified segments are university, teachers and moms. Mr. Trudeaus’ list starts with science, students and arts.

Perhaps more interesting is which segments are common and which are unique. Segments unique to Mr. Harper include moms, sports dads, gamers, polisci students, music & entertainment, and India. For Mr. Trudeau, that list includes science, arts, political enthusiasts, Vancouver, Calgary, Edmonton, Europe and Africa, and family. They share sports fans and Quebec.


Among Mr. Harper’s segments, polisci students exhibited the greatest growth (32%), followed by Toronto services (18.5%) and teachers (16.7%). His music and entertainment category shrunk (-0.7%).


Mr. Trudeau’s biggest gains were among students (42.5%), moms and girls (33.9%) and Calgary (27.8%). In fact, he also experience significant growth in his Vancouver (26%), arts (25.7%), Edmonton (24.9%), and Toronto and Florida (21.5%) segments. His family segment shrunk (-0.5%).


Twitter audiences listen more, contribute less

An interesting part of the analysis using is being able to determine how active a Twitter user’s following is. In most cases, Twitter followers are more likely to observe and listen more than they are to jump in to the fray. This could be because people are curious or just more interested in getting their content fix than offering their own thoughts. defines a “Lurker” as a Twitter user who issues fewer than 15 tweets per month. This leads to the revelation that just over 1% of Mr. Harper’s Twitter following and just over 2% of Mr. Trudeau’s following are actively tweeting users. While that suggests a small number, it’s worth noting that as of October 22, that accounts for 12,669 and 17,969 tweeters, respectively.


The online discussion was mobile

If you were a doubter about mobile, the following chart should be a wakeup call for you. It shows that followers of Stephen Harper and Justin Trudeau who contributed to the Twitter discussion almost overwhelmingly did so using mobile devices. The following graph compares the top 13 apps used to issue tweets by the followings of both leaders. In both cases, 8 of the top 13 are directly tied to smartphones or tablets.


Participation skewed male

It’s pretty typical that Twitter activity related to Canadian politics skews male. However, women have become more involved in the online discussion. I wrote about the increasing contribution of women in March 2014, noting the gender skews had shifted from 71% male/29% female in 2012 to 63%/37% in 2014.

Participation in election shows this trend is continuing, as women accounted for 39% of the Twitter activity. There were five weeks of the campaign during which women accounted for 40%. The two weeks of the Duffy trial saw an increase in male participation (64%/36%).elxn42-gender

I did not conduct gender analysis by issue. If I find myself blessed with a week of nothing to do, perhaps I’ll immerse myself in such an analysis.

Ontario led the country, Canada led the world

Ontario accounted for 49% of relevant Canadian Twitter activity over the course of the election, BC followed at 18%, Alberta 14% and Quebec 10%. Quebec commanded a 15% share the week of the TVA French language debate, and 14% the week after. BC’s strongest week was October 4 through 10 (20%) while Alberta peaked at 16% during the second and third weeks of the campaign. Saskatchewan’s strongest showing was 4% during the week of September 6 through 12.

Canada accounted for 89% of all relevant traffic while the US claimed a 5% share. The rest of the traffic is from around the world.