Customer Perspectives on Hotels in Bengaluru: A Machine Learning-Based Sentiment Analysis of Online Reviews

Abstract: In today's hospitality industry, understanding customer sentiments is essential for enhancing service quality and guest satisfaction. This study focuses on analyzing customer experiences expressed in online reviews of hotels, particularly on travel and hotel booking websites. Sentiment data were collected from after-service surveys and third-party platforms like Booking.com, covering feedback from 131 hotels in urban Bengaluru. Using Python and R, natural language processing techniques and machine learning algorithms were applied to categorize reviews into positive, negative, and neutral sentiments. The research follows a mixed-method approach, combining quantitative sentiment analysis with qualitative thematic insights. Data preprocessing ensured accuracy by eliminating irrelevant and duplicate entries. Tools like NLTK and TextBlob facilitated sentiment classification, while visualizations such as Folium maps and word clouds illustrated sentiment distribution across hotel locations. The study also explored correlations between customer ratings and expressed sentiments, applying techniques like ANOVA to examine variations across hotel star categories. Despite challenges such as data reliability and potential biases, findings offer valuable insights into areas of satisfaction and concern. Ultimately, this research aids hotel managers in improving service delivery, addressing customer pain points, and designing strategies that elevate guest experiences based on sentiment-driven analysis. Keywords: Customer Sentiment, Sentiment Analysis, Hotel, Machine Learning Algorithms, Customer Satisfaction JEL Classification Number: C88, D12, M31