“Getting from data to observations:
a computational perspective”
The availability of massive social media datasets presents new opportunities for the study of society, culture, and history. In order to render these datasets useful, we need ways of transforming these mountains of, often unstructured, data into well-defined observations of the human phenomena that produce them (e.g., from tweets to the temporal dynamics of political discourse). In this talk, I’ll discuss my past and continuing efforts to use computational techniques to perform such transformations, looking at a diverse array of phenomena including population demographics, migration, narrative, and hateful speech. In addition, using case studies from my work on migration and narrative, we will explore how the model-building activi-
ties involved in such efforts can guide the reformulation of the original research questions that lead to new phenomenological insights.
Derek Ruths is Associate Professor of Computer Science at McGill University.