Sri Lanka's Atypical Election: Analyzing the Political Discourse

sri-lanka-election
On January 8th, 2015, Sri Lanka elected Maithripala Sirisena as its new president in what media outlets across the globe considered a shock victory over the incumbent President, Mahinda Rajapaksa. The off-cycle election had been surprisingly called for by Mr. Rajapaksa on October 20th, 2014 and the official nominees were not confirmed until December 8th, only a month before polls opened. Despite these factors, 81.5% of registered voters turned out for the election, a Sri Lankan record.
Here at Crimson Hexagon, we naturally asked: could social media have played a role in voter turnout and, in turn, the result?
Setting a Buzz Monitor to track the social media surrounding the election from the October 20th announcement through the January 8th election date, ForSight aggregated over 900,000 posts across all languages referencing the candidates, the election and the most common handle tied to the election, #PresPollSL. Using the Geography feature from these results we were able to see that Sri Lankans were engaged in the election on social media[1]. Our Buzz Monitor indicated that 80% of the data was categorized as Neutral, which included news shares and reposts of official announcements while the remaining 20% was roughly split between Positive and Negative.
sri-lanka-auto-sentiment
This Buzz Monitor provided a good overview of the social media activity surrounding the event. But, were there additional sentiments behind the Positive & Negative discussions? Did some of the news shares categorized as Neutral favor a certain candidate? Clearly, deeper insights beyond a basic overview  were needed, so we decided to run an Opinion Monitor to utilize the powerful machine-learning features of ForSight.
The same keywords were used when setting the Opinion Monitor. To delineate categories for the monitor, some further secondary research was done on the election, the candidates and the political outspokenness of Sri Lankans. Additionally, we discovered that Mr. Sirisena did not have an official Twitter handle or social media campaign. By contrast, President Rajapaksa had employed Arvind Gupta, the architect of India’s recently elected Prime Minister Narendra Modi’s social media campaign. Given the information we’d gathered & knowing the outcome of the election, we set a specific goal of finding the sentiment behind social media support for Mr. Sirisena using these parameters:

  1. Posts that portrayed negative sentiment toward and/or shared negative news about the current regime were trained POSITIVE for Mr. Sirisena
  2. Posts that alluded to stability in the current regime, positive views of of the current regime or negativity toward Mr. Sirisena or the challenging parties were trained NEGATIVE.
  3. Concerns about ethnic tensions, potential violence or past violence were trained into a separate category by identifying posts that contained associated keywords & articles.
  4. Other current events, such as fishing controversy, human rights, cricket or Sri Lanka in general were trained to be OFF-TOPIC.


Example of Parameter 1: a “neutral” new share tweet is identified as “positive” in the monitor because it links to an article with positive sentiment for Sirisena.
After training the ForSight platform with 20+ posts from each category, our BrightView sentiment platform tells us that Positive Sentiment for Mr. Sirisena was a whopping 66% while Negative Sentiment was only 6%.
sri-lanka-brightview
Moreover, we are currently know from our ForSight Geography analysis that Sri Lanka is exhibiting the highest posts per capita on a world view. But, what if on the Sri Lanka Election monitor results we were able to look at voting regions that Mr. Sirisena carried[2] on the map below compared to areas of high social media activity in Sri Lanka[3]? We’ll be releasing an update to this blog to take a more granular look at this and other data surrounding our monitor.
sri-lanka-world
This is just a snapshot of the insight that ForSight’s 20+ tools can provide. Yet, this brief study sheds light on the impact that social media had on this historic campaign and makes us wonder whether Maithripala Sirisena’s victory was as much of a shock as has been reported. A full analysis utilizing native speakers, political experts & advisors would further unlock the power of our machine-learning platform and yield answers to strategic questions surrounding this event.
[1] Official Election Results: http://www.srilankanelections.com/results/results-main.shtml Map: ***http://upload.wikimedia.org/wikipedia/commons/4/4b/Wahlbezirkskarte_Praesidentschaft_Sri_Lanka_2015.svg
[2] Northern & Eastern regions carried by Sirisena with little or no data are largely rural and/or inhabited by the Tamil ethnic minority that had previously sworn allegiance to his campaign.

Request a Demo

Ready to transform your business?

Get a walkthrough of Crimson Hexagon and learn how consumer insights can help you make better business decisions.