* Sentiment Analysis of Opioid Tweets
Brief details
For my senior capstone project at the University of Rochester, I collaborated on a health information data science project sponsored by a professor specializing in social media data. The project focused on analyzing Twitter data to investigate the relationship between opioid use, particularly among users of medication-assisted treatments (MAT), and dental health issues. Given the rise of opioid misuse and its associated health impacts, our goal was to provide insights into public perceptions and geographic trends, as well as uncover any patterns related to dental issues.
Summary
- Constructed and implemented data pipeline, extracting and cleansing 3M tweets using the Python NLTK package
- Developed BERT-based classification model via Hugging Face library to classify tweet context, reach precision of 85%