Customer Sentiment Analysis on OTA Platforms: Insights for Enhanced User Experience and Service Optimization

Tourism is a key driver of global economic growth, contributing 9.1% to the global GDP in 2023. As a renowned archipelago with over 17,000 islands, Indonesia has become one of the world's top tourist destinations, with tourism contributing 3.6% to its national economy in 2022. Evolving technology also brough digital transformation to this industry, part of it is online travel agency (OTA). Customers feedbacks are something app developer company encounters everyday. With huge numbers of users, number of feedbacks compiled over the times are getting larger every single day. Sentiment analysis can be an alternative solution to address complaints as a priority, enhancing user experience through optimized customer service from the company. Utilizing IndoBERT representation and long short-term memory (LSTM) for sentiment analysis classification (IndoBERT-LSTM), this research aim to optimize OTA platforms service for better customer experience. This research includes data collection trough scraping from google playstore review on Traveloka, with a few data cleansing, ensuring consistency within the data. Leveraging IndoBERT for a contextual representation and LSTM to classify customer feedbacks into respective sentiment category. This model successfully achieved a high performance with an F1Score of 93.1% and 84.34% recall on the negative class (class 0), making the model adequate enough to identify complaints and discomfort expressed by users in the review section in Google Play Store.