Fun in the Philippines: Automatic Identification and Sentiment Analysis of Tourism-Related Tweets

With the growing use of social media in the Philippines, tourism-related user-generated content is readily available. As a growing hub of tourism and culture, this could be particularly useful to the country. However, a large amount of this data has gone unanalyzed. This study discusses and develops a way that could help bridge that gap using automated tourism-related tweet identification with Support Vector Machines and sentiment analysis with Naïve Bayes. F-scores of 0.943 and 0.81 were obtained by these components respectively, with the overall system obtaining an accuracy of 84%. Mapbox was used for visualization, with tweets plotted based on their geolocations and sentiments. This study can be used as a way of gathering tweets from the Philippines, identifying which could be relevant in terms of tourism information and presenting these in a way that could be useful and easy to understand and interpret.