Online User Reviews Investigation Towards Madura Island Tourism using Latent Semantic Analysis

Modern tourism relies on customer experience feedback that is often shared via social media in images, videos, or texts. The massive growth of user-generated content, especially text, opens the opportunity for research to help us analyze and extract the information from the documents. Natural Language Processing methods enable us to extract data from users’ opinions in unstructured text, which is helpful for business strategy making. In this study, we summarize the terms often discussed in the Google Maps review regarding Madura Island using the topic modeling technique. One well-known topic modeling method is Latent Semantic Analysis, an unsupervised learning method that extracts the documents based on the appearance frequency of words. This method then calculates the distance between words and clusters of similar terms. By extracting 11240 reviews from Google Maps, we obtain ten words representing the themes that often emerge in user reviews. The emerging themes then are beneficial for further analysis to decide the business development strategy for the tourism sector.