Leveraging NLP to Uncover Hidden Insights in Stakeholder Feedback

This study investigates the use of natural language processing (NLP) approaches to improve stakeholder feedback 
analysis, with an emphasis on sentiment analysis, clustering, and engagement prediction. The study used a dataset obtained from 
several sites, including Twitter and Yelp, and found that 65% of the reviews studied were favorable, indicating that users were 
generally satisfied. Exploratory data analysis (EDA) revealed that typical phrases connected with favorable sentiments. The study 
used three machine learning models to predict sentiments: Logistic Regression, Random Forest, Neural Network (NN) and Support 
Vector Machine (SVM), with Logistic Regression attaining the maximum accuracy of 95%. The findings emphasize the crucial 
necessity for enterprises to resolve service concerns, especially on Mondays, when negative feedback is high. By integrating data
driven insights from sentiment analysis and machine learning, this study highlights the potential for enterprises to effectively 
improve customer engagement and operational strategy. The report finishes with recommendations for improving customer service and proactive sentiment monitoring in order to develop stronger stakeholder relationships.

IndexTerms - Stakeholder feedback Analysis (SFA), Natural Language Processing (NLP), Exploratory Data Analysis (EDA), 
Support Vector Machine (SVM), Cluster 0 (Positive Reviews), Cluster 1 (Neutral Reviews), Cluster 2 (Negative Reviews)