Sentiment analysis for customer review: Case study of Traveloka

Abstract

This study will discuss customers’ satisfaction with the services of Traveloka by analyzing how many people are satisfied and unhappy with the services that Traveloka has to offer. This study uses Twitter to acquire all the data we need, focusing only on tweets about Traveloka. The dataset is gathered from Twitter API, which consists of 1200 tweets related to Traveloka. Scikit-learn library is used through python to do the analysis process. This research employs three classification metheods: Support Vector Model (SVM), Logistic Regression, and Na¨ıve Bayes. The steps in this research were data retrieval, transformation, classification training and predicting the test data, and finally, the result analysis. Therefore, this research is looking forward to how most Twitter users feel about the performance of this mobile traveling application. The result shows that SVM has better accuracy in determining the sentiment of tweets about Traveloka.