CONSISTENCY EVALUATION OF HOTEL RATINGS

In Hotels industry, reviews and ratings have become a critical aspect of their business, as staggering 97.7% customers read online reviews from other travelers brings attention to the surge of review sites. But, with so many travelling agencies facilitating booking services and reviews of their own, Does there exist a consistency between them? Do they represent similar values and ratings? Does customer reviews are reflected in the ratings for hotel? To answer these questions, we did web analysis for consistency evaluation of ratings and reviews of hotels across different travelling agency websites.

We selected three different websites Booking.com, Expedia.com and Hotels.com, who have their own rating criteria for hotels. We selected a total of 30 hotels across New York city and to expand horizon of our research, we selected 10 hotels each for the three categories 5 Star hotels, 3 Star hotels and Unrated hotels. We crawled the ratings and reviews of these hotels across all the three websites. The ratings across these three websites were normalized to single scale and criteria for further analysis.

We crawled the reviews for this hotels using Python and by taking advantage of Selenium Web Driver and chrome driver. We performed the sentiment analysis on the reviews crawled, by first grouping the words in reviews under 5 topics – Cleanliness, Location, Comfort, Condition and Staff. Using another list of adjectives under good and bad, we determined the polarity of the reviews across the five topics and see if these sentiment was reflected in the ratings of the hotels across these different websites.

Sentiment Analysis result of reviews of an Unrated Hotel

From our study we found that 5-star hotels show good consistency as expected, 3-star hotels show high consistency for overall ratings but sub-ratings having inconsistency and Unrated hotels show high inconsistency overall. People care a lot about Room Comfort and cleanliness, when it comes to any hotel category and same is reflected in their reviews. They have high complaints and varied opinions about the room comfort and speak lot of good things about the Cleanliness and Staff and have similar opinions. Overall, sentiment analysis shows reviews are reflective of the ratings for these hotels.

Below is the video of project presentation for your reference. This was a team project and my team members were Maggie Zhu, YuanYuan Pei and Yuyi Zhang.