Analisis Sentimen Pelanggan Kopi Kenangan pada Media Sosial Instagram Menggunakan Metode Lexicon Based

Abstract

This study aims to analyze customer sentiment towards Kopi Kenangan on Instagram using a lexicon-based method. The rapid growth of Kopi Kenangan as a local coffee brand in Indonesia has made Instagram a key platform for building brand image and interacting with customers. The main problem faced is the large volume of unstructured comment data that cannot be processed manually, so an efficient and systematic automated approach is needed. This study uses a case study methodology and quantitative descriptive techniques. The object of research in this study is the official Kopi Kenangan Instagram account, and the dataset used to conduct this study comes from that account. The dataset that can be used to test this research will be created by developing processing stages. The Lexicon-Based method is used in this research approach. The purpose of the dictionary-based Lexicon-Based method is to determine the weight of sentences in the dataset to identify sentiment class labels. The data used in this study comes from 1689 records that have been preprocessed to produce 1438 patent records by removing empty comment records and verifying duplicates up to a certain threshold. The next step is to identify comments based on their sentiment: 25 negative, 1347 neutral, and 317 positive. The results show that negative sentiment reached 1,5%, neutral 79.8%, and positive 18.8%. Based on this presentation, the majority of the text examined was consumer responses to Kopi Kenangan's Instagram posts, and most of them were negative. This indicates that the Lexicon-Based model developed is capable of classifying sentiment with good accuracy.