Research Assistant (doctoral research) - tweet dataset coding and stance labeling
Don't disclose
Text
Classification
TextClassificationAssisted in doctoral research by coding a political tweet dataset and categorizing each tweet’s stance on abortion. The work involved mapping tweets to discrete labels: “support,” “oppose,” or “neutral.” You also supported comparison of human-coded annotations against a proprietary computational processing system. • Labeled tweet stance using a multi-class classification scheme. • Performed evaluation of alignment between human annotations and model outputs. • Organized and prepared labeled data for downstream analysis. • Supported validation of a proprietary system’s performance relative to human judgment.
2023 - 2024