Mining Tourists’ Opinions on Popular Indian Tourism Hotspots using Sentiment Analysis and Topic Modeling

Abstract

User-generated content is an exploration area of interest with regards to web 2.0. The development of social networks and community-based websites have changed the manner in which individuals utilize the Internet. It makes individuals no longer restricted to pursuing the data given by professional channels, but to making individual profiles, producing personalized content, or sharing photographs, recordings, blogs, and so forth. This sort of data comprises the current online user-generated content. With the continuous development of the travel industry, the quantity of online travel review websites has also increased. Indian Tourism is popular for its rich culture and diversity and hence Government of India has increased the number of new tourist destinations to expand their popularity and presence. Researchers have proposed various studies to increase tourism network using Big Data. Techniques of Sentiment Analysis along with Topic Modelling have been used to unearth patterns and observations from online reviews. This paper aims to mine reviews of 10 popular travel destinations in India. Using sentiment analysis technique, the proposed research work has explored the polarity of various reviews extracted from TripAdvisor. Data collection was done by using the web framework Scrapy to acquire more than 10,000 reviews for these destinations. This paper also analyzes the result of doing Topic Modeling on reviews for individual destinations. Results conclude that Joy is the most common emotion in all the visitor’s experiences. Indian tourism decision quality can be improved by the help of the results from this study.