My recent projects are focused on measuring influence on Twitter. In particular, I’m measuring changes in topical attention and expressed topical sentiments after exposure to suspicious content, such as bots, misinformation, and coordinated information campaigns. I am also working on developing new tools necessary for this task, especially Deep Learning models and Natural Language Processing (NLP) systems to do keyword extraction and targeted sentiment analysis (a.k.a. aspect-based sentiment analysis, or ABSA), and Multivariate Hawkes Process models for causal inference.
I am applying some of these measurement tools in the CoVaxxy project in which we are studying COVID vaccine hesitancy, and its relationship to COVID public health behaviors and outcomes.