Drugsideeffect: A Python Package for Exploratory Pharmacovigilance from Unstructured Text

Abstract

Drugsideeffect is a Python package for exploratory pharmacovigilance using unstructured health-related text. It differs from general-purpose NLP libraries by integrating supervised filtering, symptom lexicons, slang normalization, temporal expression extraction, alert keyword detection, sentiment analysis, and visualization into a single workflow specifically designed for exploratory pharmacovigilance. Informal user-generated narratives are transformed into structured drug-related signals for the analysis and visualization of drug side-effect discourse in unstructured text. The package supports preprocessing, symptom detection, alert keyword identification, temporal analysis, sentiment scoring, and visual analytics. Designed for research and education, Drugsideeffect enables exploratory pharmacovigilance studies using social media, surveys, and other unstructured health-related text.