List All Projects   List Graduate Projects   List Senior Design Projects   List REU Projects   Other Undergraduate Projects

Project # 1
Aspect Based Sentiment Analysis(Diversity and Inclusion)

Team Lead: Geetha Madhuri Sigatapu
Other Team Members:
Faculty Advisor(s): Mengqi Zhan
Department: CSE
Project Type: Graduate Student
Title: Aspect Based Sentiment Analysis(Diversity and Inclusion)
Abstract: This project analyzes S&P 500 company reviews to uncover insights related to Diversity, Equity, and Inclusion (DEI). As organizations prioritize inclusive work environments, understanding employee perceptions is crucial. The project involves developing a pipeline to clean, process, and analyze textual data for actionable insights. The dataset, sourced from employee reviews, undergoes preprocessing to remove duplicates, handle missing values, and filter DEI-relevant content using keywords like ”diversity,” “inclusion,” and “equity.” Reviews lacking meaningful DEI context are excluded to maintain focus. Advanced Natural Language Processing (NLP) techniques, including TextBlob, BERT, and OpenAI GPT, are applied. TextBlob provides initial sentiment scores (polarity and subjectivity), while BERT offers fine-tuned binary sentiment classification with contextual depth. OpenAI GPT enhances analysis through context-based scoring for DEI relevance and employee perception intensity using custom prompts and few-shot learning. The outputs are consolidated into a comprehensive dataframe with DEI relevance labels, keyword scores, and sentiment ratings. Discrepancies across models are resolved through averaging, while consistent results reflect consensus. The enriched dataset is exported as a CSV for further analysis and visualization. The outcome is a validated DEI analysis framework, offering organizations data-driven insights to guide inclusivity strategies, highlighting the value of precise data preparation and multi-model NLP approaches.
Poster: View Poster #1 (PDF)
Judges Scoring: Submit Evaluation (judges only)