Automated content analysis
Machine learning pipelines over large corpora of political and media text for quantitative social science research at the CLESSN, evaluated against human-coded samples.
The CLESSN, a political science research chair at Université Laval, studies political and media discourse: far more text than a team could read by hand. I built the machine learning pipelines that made that possible: collecting large corpora of political and media text, classifying and measuring their content automatically, and handing researchers data they could analyse quantitatively.
Every model was evaluated against samples coded by humans, so the results could be checked.
I also supported and trained the graduate students who used these tools.