Perak's tourism through the lens of social media: a computer-based sentiment analysis approach / Raudatul Jannah Rostam, Azilawati Azizan and Nurkhairizan Khairudin

• Main Authors: Rostam, Raudatul Jannah, Azizan, Azilawati, Khairudin, Nurkhairizan
• Format: Article
• Language: English
• Published: Faculty of Hotel & Tourism Management, Universiti Teknologi MARA 2023
• Subjects: Travel and state. Tourism
• Online Access: https://ir.uitm.edu.my/id/eprint/94870/1/94870.pdf; https://ir.uitm.edu.my/id/eprint/94870/; https://www.jthca.org/
• Institution: Universiti Teknologi Mara

Perak, a state in Malaysia, has a lot of exciting and captivating places for tourists to visit. There are many different natural wonders, cultural landmarks, historical sites, and delicious foods to enjoy. This makes Perak special and can bring in a lot of tourists. In today's digital age, social media plays a big role in sharing information, including tourism. With the presence of technology that can discover feelings and emotions from social media texts, such as sentiment analysis, we can make this even better. Sentiment analysis is the process of analyzing and identifying the feelings conveyed in a text such as positivity or negativity by utilizing natural language processing (NLP) and machine learning approach. This project aims to discover the sentiments of tourist attraction in Perak by analyzing Twitter data. The project has three objectives: first to collect and prepare a suitable and reliable dataset, then classify the data into positive, negative, or neutral sentiments using NLP techniques, and finally develop a web-based application to visualize those sentiments. To accomplish these objectives, a collection of tweets pertaining to Perak’s tourist attractions has been gathered and prepared for analysis. TextBlob library in the Python programming language is used to extract sentiment of tweets from Twitter data and classify them into positive, negative, or neutral categories. Then a machine learning approach, Support Vector Machine (SVM) is used to create, train and test the sentiment model. And as a result, utilizing the SVM classifier with a linear kernel and a split of 70:30 between training and testing data yields an increased accuracy rate of 75.50%. This project is important because it provides valuable insight to the tourism sector in Perak.