BUS5CA Customer Analytics and Social Media: Sentiment Analysis Report

This report presents a comprehensive sentiment analysis of hotel review tweets, fulfilling the requirements of a BUS5CA assignment. The analysis employs both dictionary-based and machine learning approaches using R programming language. The preprocessing steps involve data cleansing (removing URLs, usernames, and special characters), tokenization, lowercasing, and stemming to create a Bag-of-Words representation. The report details the development of sentiment analytics engines and applies them to predict the sentiments expressed in hotel reviews. Furthermore, the analysis incorporates the use of SAS Sentiment Analysis Studio to build a statistical model, evaluate its accuracy, and compare results with the R-based models. The report provides insights into the relationship between sentiment, keywords, and potential business applications, demonstrating a practical application of data science techniques in understanding customer preferences and social media analysis. The student has used XGBOOST classifier algorithm to get accuracy and ROC curve is also used to evaluate the model. This report is available on Desklib, a platform offering AI-based study tools for students.