Solo travel is a growing trend in leisure travel. More people are embracing the idea of seeing the world on their own and more facilities, such as hostels, are increasing to support that. In this project for Computational Research course, Dhruval Bhatt, applies n-gram analysis on text from reviews of popular hotel and hostel in San Francisco to analyze how the response varies between solo travelers and other types of travelers. Then a decision tree model is applied on descriptive text as predictors and travel type as the predicted variables. The exploratory text analysis showed that different traveler types do in fact value different things in a hotel/hostel. Solo travelers value staff and guest friendliness more while business travelers appreciate comfort and amenities. The model had a 62.1% accuracy for prediction.
Class: MACS 30200/1 - Perspectives on Computational Research at the University of Chicago
Professor: Dr. Benjamin Soltoff
Term: Spring 2019
- Data code file for scraping reviews and manually compiled TripAdvisor Review data
- Code All code files for data prep, exploratory analysis and predictive model
- Project Proposal
- Final Project Report
- Poster