NLP-Driven Sentiment Analysis for Stakeholder Feedback in Sustainable Projects

This study applies advanced Natural Language Processing (NLP) techniques for sentiment analysis to interpret stakeholder feedback in sustainable project management. Stakeholder opinions on sustainability-focused projects, which address environmental, social, and economic factors, are critical for aligning projects with community needs. Using a hybrid approach that integrates Support Vector Machines (SVM), Long Short-Term Memory (LSTM) networks, and the VADER lexicon-based model, this research automates the analysis of textual feedback to reveal sentiment trends and key topics. The methodology includes extensive preprocessing and topic modeling via Latent Dirichlet Allocation (LDA) to identify recurrent themes such as environmental impact, resource management, and social responsibility. Results demonstrate that machine learning-based sentiment analysis, particularly LSTM, achieves higher accuracy in nuanced language contexts, compared to lexicon-based approaches. This work highlights NLP's role in sustainable project management by providing actionable insights into stakeholder priorities, enhancing data-driven decision-making, and identifying areas for strategic adjustments. Future work focuses on improving model robustness, especially in handling multilingual and informal data for more comprehensive stakeholder sentiment analysis.