This project presents an in-depth Aspect-Based Sentiment Analysis (ABSA) of customer reviews, particularly focusing on the hotel industry. Utilizing Python's powerful libraries, the analysis delves into trigram frequency patterns, extracting crucial information about various hotel aspects such as rooms, staff, and amenities. By breaking down reviews into trigrams, we ensure a nuanced understanding of sentiment that is more granular than traditional sentiment analysis techniques.
The cornerstone of this project is the intuitive network graph visualization. This visualization represents word co-occurrence within reviews, spotlighting the most frequently discussed topics in a visually engaging manner. This approach not only deciphers the sentiment surrounding specific hotel-related aspects but also provides insights into how different aspects interrelate in customer feedback.
To get started, refer to the codebase and follow the comments for a step-by-step guide on the ABSA process.