This is a research project to analyzes the impact of sustainability related topics at some TripAdvisor hotels. The results will be published in two papers answering the three main research questions and deployed on a website that shows all the hotel summaries and ratings. In this repository, you will find all the scripts used in this research.
- User query-based system able to summarize client reviews and answer questions based on TripAdvisor comments in order to take a well-informed decision.
- Do luxury hotels reviews include more/less (and with more positive/negative sentiment) comments on sustainability related topics compared with other types of hotels?
- Is it possible to rate hotels based on their reviews content (sentiments)?
| Task | Model | Evaluation |
|---|---|---|
| Review summarization | T5 | Rouge, Bleu |
| Clustering and n-grams (topics) | BERTopic | ❌ |
| Sentiment analysis and rating prediction | VADER, HIV4, LM, NRCLex | ❌ |
