A cloud-based tool for sentiment analysis in reviews about restaurants on TripAdvisor

Abstract

The tourism industry has been promoting its products and services based on the reviews that people often write on travel websites like TripAdvisor.com, Booking.com and other platforms like these. These reviews have a profound effect on the decision making process when evaluating which places to visit, such as which restaurants to book, etc.

In this contribution is presented a cloud based software tool for the massive analysis of this social media data (TripAd-visor.com). The main characteristics of the tool developed are: i) the ability to aggregate data obtained from social media; ii) the possibility of carrying out combined analyses of both people and comments; iii) the ability to detect the sense (positive, negative or neutral) in which the comments rotate, quantifying the degree to which they are positive or negative, as well as predicting behaviour patterns from this information; and iv) the ease of doing everything in the same application (data downloading, pre-processing, analysis and visualisation).

As a test and validation case, more than 33.500 revisions written in English on restaurants in the Province of Granada (Spain) were analysed.