Signal in the network: public opinion polling using social media data

Can social media be used to make valid claims about public opinion? Skews in the usership of platforms such as Twitter and Facebook mean that the opinions of social media users usually do not reflect those of the population as whole. This project aims to develop a technique that corrects for the bias in social media data so as to make continuous, real-time, representative inferences about a target population. This project serves a case study focusing on the ability to predict political and electoral outcomes from social media. It aims to advance the scholarship in political methodology and contribute substantially to the research on political behaviour and electoral politics. It will also further the partner organization’s facility in predictive analytics using Big Data.

Faculty Supervisor:

François Gélineau

Student:

Partner:

Vox Pop Labs Inc

Discipline:

Sociology

Sector:

Professional, scientific and technical services

University:

Université Laval

Program: