Executive Summary

Airbnb is one of the top online platforms for accommodations for tourists and travellers worldwide, and according to iProperty Management, Tokyo, Japan is the most popular city for booking experiences. This study aims to determine themes that can be extracted from Airbnb listings in Tokyo using unsupervised clustering, and to see what made Airbnb so successful in the city.

Data was obtained from the public Inside Airbnb dataset and stored in an SQLite database. Term Frequency-Inverse Document Frequency (TF-IDF) for feature extraction and Latent Semantic Analysis for dimensionality reduction were applied to the corpus.

We identified 4 clusters in Tokyo’s Airbnb listing summaries: (1) Facility type, (2) Accessibility, (3) Proximity, and (4) Room Quality — reflecting what hosts tend to highlight about their accommodations. This information would be helpful to tourists, local and foreign tourism bodies, and business owners in making travel and tourism decisions.

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