The 2015 Canadian Election on Twitter: A Tidy Algorithmic Analysis

Appendix containing the robustness checks.

William Sanger https://www.cirano.qc.ca/en/community/directory/view/2109 (CIRANO and Polytechnique Montreal)https://www.cirano.qc.ca , Thierry Warin https://www.nuance-r.com/principalInvestigator.html (SKEMA Business School (Raleigh, NC))https://www.skemagloballab.io
08-22-2019

Abstract

During the 2015 General Election in Canada, the Liberal Party of Canada was elected to form a majority government, despite being third at the start of the electoral campaign. How was perceived the incumbent party and its rivals during the election? By using a tidy approach of a massive dataset collected on Twitter (3.5 millions tweets), we developed two methodologies to characterize how politicians were perceived on social media during the election. First, a sentiment analysis was performed regarding each political leader, then by using the whole dataset, different topics of the election were associated to each leader through a Latent Dirichlet Allocation (LDA).

Keywords: Social Media Analysis, Elections, LDA, Social Data Science,Canada

Citation

For attribution, please cite this work as

Sanger & Warin, "SKEMA Global Lab in AI: The 2015 Canadian Election on Twitter: A Tidy Algorithmic Analysis", IEEE Xplore, 2019

BibTeX citation

@article{sanger2019NigeriaElection,
  author = {Sanger, William and Warin, Thierry},
  title = {SKEMA Global Lab in AI: The 2015 Canadian Election on Twitter: A Tidy Algorithmic Analysis},
  journal = {IEEE Xplore},
  year = {2019},
  note = {https://skemalab.io/posts/2019-08-22-the-2015-canadian-election-on-twitter},
  doi = {10.1109/CSCI.2017.158}
}