Eyesover Technologies

​  BEHIND THE NUMBERS.

  • Home
  • Research
  • Blog
  • About Us
    • Privacy
  • Contact
  • Log in
  • Careers

9/30/2015

The Munk Debate on Canada’s Foreign Policy

Read Now
 
Picture
​As many Canadians were watching the Blue Jays move ever closer to the American League East pennant Monday night, those who prefer politics to pennant races watched what was by most accounts, an equally entertaining spectacle.

Of course, we know who won the Jays game, but the winner of the Munk Debate is still being debated across the nation.

Here at Eyesover, we followed the debate Monday night and while we recognize the “winner” of a debate is a fairly subjective call, our system did produce some interesting insights.

Before looking at the results, it is important to understand what we do at Eyesover. Our system gathers online data from social media pertaining to social issues our customers care about. We then analyze the data using various metrics that provide more insight into what the data is telling us. One basic metric we use is sentiment analysis, which determines if comments about a certain subject are positive or negative. Measuring sentiment towards specific political issues is not the same as measuring voting intention, but it can provide excellent insight into public opinion on a subject or issue using a massive amount of data.

Our sentiment analysis is based on a scaled range of 0 (negative) to 100 (positive) with 50 equating to a neutral sentiment. Keep in mind that we find extremely positive or negative comments are rare with most tending towards neutral.
​
The first measurement we took was the average sentiment score for each of the leaders from all comments in which they were mentioned. We analyzed over 37,000 online comments pertaining to the debate and found Justin Trudeau had the highest sentiment score at 74.06 followed by Thomas Mulcair at 63.93 and Stephen Harper at 58.17. These scores indicate that on average, comments about all of the respective leaders were relatively positive, a finding that mirrors much of the post-debate analysis where there was general agreement all three leaders performed well. 

Picture
​The second measurement was the average sentiment of each individual commenting. If you follow online discussions, you know this is important due to the fact that some individuals will post numerous comments that can skew an overall average if they are extremely negative or positive towards the subject. Not that anyone would try to skew the public perception in this manner during an election campaign of course…
​
When we count the number of individuals (n=14241) who posted the comments that we analyzed, we learn the average comments per individual was approximately 2.6 with many posting once while some commented over 100 times. Giving equal weight to each individual’s average sentiment still produced similar results for each leader - Justin Trudeau at 71.80 , Thomas Mulcair at 62.31, and Stephen Harper at 59.81 - but we see that removing the impact of multiple posts from one individual increased the score for Harper (+1.64) and decreased the score for Trudeau (-2.26) and Mulcair (-1.62).

This highlights a problem with using simple metrics such as mentions in determining an outcome of an event based on online discussions. We found that individuals posting numerous comments about Stephen Harper were primarily negative while those posting multiple times mentioning Justin Trudeau and Thomas Mulcair were primarily positive. For example, the individual who commented on Stephen Harper the most commented 142 times with an average sentiment score of 46.79. This compares to the leading Thomas Mulcair commentator who posted 94 comments with a score of 79.75. This provides a good reminder about the need to dig deeper into the data to truly understand what is taking place online.
​
Overall, the online community had positive things to say about all three leaders, but at the end of the evening, those discussing Justin Trudeau were the most positive.

Share


Comments are closed.
Details

      Subscribe

    Subscribe to Blog

    Archives

    October 2019
    September 2019
    August 2019
    July 2019
    June 2019
    May 2019
    April 2019
    March 2019
    November 2018
    July 2018
    June 2018
    May 2018
    April 2018
    March 2018
    February 2018
    December 2017
    November 2017
    October 2017
    May 2017
    March 2017
    November 2016
    September 2016
    August 2016
    July 2016
    March 2016
    February 2016
    October 2015
    September 2015

    Categories

    All
    Artificial Intelligence
    Enterprise
    Eyesover
    Eyesover News
    Political Polls
    Political Software
    Predictive
    SaaS
    Social Media Monitoring
    Trend Discovery

Company

About
Location



Support

Contact
Privacy Policy
© COPYRIGHT 2015. ALL RIGHTS RESERVED.
  • Home
  • Research
  • Blog
  • About Us
    • Privacy
  • Contact
  • Log in
  • Careers