Executive Summary

Tourism is one of the drivers of New York City’s economic prosperity — in 2019 alone, 66.6 million visitors spent over $46.4 billion USD on hotels, restaurants, and other establishments across the city’s five boroughs. One of the city’s catalysts in this success is the rise of Airbnb, the top online platform for accommodations for tourists and travellers worldwide. According to iProperty Management, New York City is the 4th most popular city for booking experiences, with more than 40,000 unique listings. With this many options for travellers, this study aims to determine which Airbnb listing in New York City can be recommended to a user.

Data was obtained from the Inside Airbnb dataset and stored in an SQL database. Since there are no available numerical ratings per user, Natural Language Processing was used to extract the polarity of each user’s text reviews and develop a user-based collaborative filtering model.

With the combination of sentiment analysis and user-based collaborative filtering, the recommender system can provide suggestions to a user based on their previous reviews, while also taking into account preferences such as room type, borough (location), and price range.

Full text article and source codes can be provided upon request.