Analysis of Public Sentiment Text Clustering on Tax Increases using Orange Data Mining on Twitter

E-ISSN : 2807-9035 Volume 5, Number 1, May 2025 https://doi.org/10.47709/brilliance.v5i1.5787 Analysis of Public Sentiment Text Clustering on Tax Increases using Orange Data Mining on Twitter Ibnu Azhar Maulana1*, Ari Purno Wahyu Wibowo2 1,2 Widyatama University, Indonesia , *Corresponding Author Article History: Submitted: 17-04-2025 Accepted: 22-04-2025 Published: 29-04-2025 Keywords: Clustering; Data Analysis; Orange Data Mining; Tax; Text Mining. Brilliance: Research of Artificial Intelligence is licensed under a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). ABSTRACT Taxes play an important role in the life of a nation and state, particularly in the implementation of national development. Recently, Indonesia issued a new policy to increase VAT to 12%. This policy has sparked a range of both negative and positive opinions from the public. As a result, various reactions and sentiments have been expressed by citizens regarding the policy. To analyze these public sentiments, text mining was carried out using the Orange Data Mining application, utilizing data from the Twitter platform to observe and evaluate Indonesian citizens' reactions. A total of 100 tweets were collected using relevant keywords to find content related to the policy. The results were then categorized into several sentiment groups based on the similarity of their content. After the text classification, the data was stored in a table showing the number of positive, negative, and neutral sentiments. This data was later visualized in a graph, which revealed that the most common reaction was disappointment, followed by confusion, enthusiasm, and lastly, anger. The results of this study indicate that many Indonesian citizens are disappointed with the VAT increase policy. Many believe that the government's use of tax funds has not been satisfactory. Therefore, the government is urged to improve its programs so that citizens can feel the benefits of the taxes they pay. INTRODUCTION The government holds the authority to establish tax collection policies as a primary means of achieving the common welfare. Although the benefits of taxes are not always directly felt by the public, taxes remain a fundamental component of state financing (Gunawan et al., 2020; Koynja, 2023). Taxes play a vital role as a fiscal instrument that enables the government to fund various development programs and public policies. Amid national and global economic dynamics, Indonesia's tax system continues to undergo adjustments to maintain fiscal stability. One strategic type of tax that supports the nation's economy is the Value Added Tax (VAT). Tax increase policies, including VAT, often spark public debate. The implementation of the 12% VAT rate, which will take effect in 2025 as stipulated in the Law on the Harmonization of Tax Regulations (UU HPP) of 2021, has triggered widespread discussion and various public responses (Al Mustaqim et al., 2024; Febiola, 2024). Some people are concerned about the impact of the increase on purchasing power, particularly for low-income groups and Micro, Small, and Medium Enterprises (MSMEs) (Yudistira, 2024). Public attitudes towards this policy are greatly influenced by how the government conveys information and the role of social media in shaping public opinion (Ardiansyah, 2022). Social media is now the primary forum for the public to voice their opinions and respond to government policies. Twitter, now also known as X, is one of the most actively used platforms in Indonesia, with the number of users reaching around 27.5 million as of October 2023, making Indonesia the fourth most active country in terms of usage worldwide (Annur, 2023). Twitter is a valuable source of data for analyzing public sentiment on various issues, including the VAT increase policy (Wahyuni, 2022). In this study, data from Twitter will be obtained through the use of the Twitter Application Programming Interface (API), which allows researchers to systematically access and collect tweets. The data will then be analyzed using Orange Data Mining software, an open-source visual platform that supports various data mining methods, including clustering. This technique is used to identify patterns and trends in the sentiments expressed in tweets by Twitter users. This study aims to analyze how the sentiment of the Indonesian people regarding the VAT increase policy is grouped based on the similarity of content and expression of opinion. The results of this study are expected to provide constructive input for the government in responding to public aspirations, as well as an evaluation of the effectiveness of public policy communication through social media. LITERATURE REVIEW Research by (Anwar & Permana, 2023) conducted sentiment analysis on electric vehicle products on the Twitter platform using the VADER analysis technique. VADER (Valence Aware Dictionary and Sentiment Reasoner) is a This is an Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. 93 E-ISSN : 2807-9035 Volume 5, Number 1, May 2025 https://doi.org/10.47709/brilliance.v5i1.5787 lexicon- and rule-based sentiment analysis tool specifically tailored to sentiments expressed on social media. In another study, (Fauziah, 2024) conducted an analysis of public perception regarding the VAT rate increase in Indonesia, which revealed that 6% of the sentiment was positive, 78% was negative, and 16% was neutral. Tax One of the most important sources of national revenue for the growth and improvement of public welfare is taxation. Taxation is a monetary obligation that must be fulfilled by individuals, businesses, or other legal entities in accordance with the law (Lestari et al., 2024). Taxes play a crucial role in a state. Every citizen has an obligation to contribute voluntarily to their country for the purposes of state development, where the funds collected from taxes will play an essential role in supporting national development and government expenditures. Orange Data Mining Orange Data Mining is a data analysis software developed by the University of Ljubljana in Slovenia. It was first released in 1997. Orange has become one of the most popular data analytics platforms in the world, mainly due to its focus on an intuitive user interface and the ability to perform a variety of data analysis tasks without the need for coding (Ferri, 2024). Orange Data Mining is currently widely used by people to analyze data or conduct data mining because the application provides the features needed to help users obtain the information they need. Text Mining Text mining is the process of discovering unknown knowledge through automatic information extraction from a large number of unstructured texts (Hermawan et al., 2023). Text mining is a method that can be used to handle classification, clustering, information extraction, and information retrieval problems (Afdal & Elita, 2022). (...truncated)