Dimas Haafizh Rahman, Zharifah Dhiya Ayu Alistin, Setia Pramana

Sentiment Analysis and Topic Modelling of Public Opinion on Statistics Indonesia Data in 2024

Introduction

Sentiment analysis and topic modelling of public opinion on statistics indonesia data in 2024. Analyze public sentiment & topics on Statistics Indonesia (BPS) data in 2024 from X tweets. Uses ML (SVM, LSA, LDA) to reveal opinions on data quality & issues like poverty rates.

Abstract

The role of BPS has become increasingly crucial with the rising demand and sources of data over time. The quality of BPS data is evaluated through the Data Needs Survey (SKD). The 2024 SKD indicates that 98.16% of consumers are satisfied with the quality of BPS data. However, this evaluation only involved data consumers from BPS PST, and there remains a time gap between the implementation and dissemination of the survey results. Social media platform X, which is popular in Indonesia, allows its users to express their opinions through tweets. This research is conducted to understand public sentiment, identify the best classification model, and discover topics discussed by the public regarding BPS data based on tweets from the X platform in 2024. The tweets were taken through labeling and preprocessing before applying Machine Learning methods to classify public sentiment. The Support Vector Machine (SVM) method, with a weighted average of 0.68, performed best compared to Naïve Bayes, Rocchio Classification, and K-NN in modeling public opinion sentiment. The implementation of LSA and LDA discovered topics consisting of public opinions and issues related to BPS data such as poverty rate manipulation and BPS data as a credible source.

Full Text

Comments

You need to be logged in to post a comment.