Abstract

Formaggio Coffee and Resto Tangerang serves western dishes with flavors that are tailored to the tastes of the Indonesian people. The increasing number of restaurants in Tangerang every year makes Formaggio Coffee and Resto supposed to have a competitive advantage by increasing customer satisfaction. Customer satisfaction can be obtained if customer expectations are met. Formaggio management considers criticism and suggestions given by customers as a positive thing that can improve their performance. However, the large number of customer reviews scattered across various sites makes it difficult for the restaurant to manage their customer opinions. This can be overcome by web scraping on the review sites, such as Traveloka, PergiKuliner, Zomato, and Google Reviews. The data that has been collected is 741 reviews with a time span from 2018 to 2021. Then, to get the information from customer reviews, sentiment analysis can be implemented using the Indonesian Sentiment Lexicon dictionary, TF-IDF features, and Multinomial Naïve Bayes Classifier. The classification model was tested using confusion matrix with four parameters, such as accuracy, recall, precision, and f1-score. The average value of each parameter is 95%, 68%, 85%, and 72%. The results of the research were visualized into a dashboard and tested using the System Usability Scale (SUS) questionnaire with 67.5 as a final score, which means the dashboard is well received by Formaggio management.