Social media analysis of BPS data availability in economics using decision tree method

Badan Pusat Statistik (BPS) is one of the government institution with the duty to provide accurate and reliable data for the government and society uses. Currently BPS does not have basic tool that can be used to evaluate usage of their data by people or government. The aim of this study is to analyze the usage of BPS data, especially in economics by people or society via social media 'Twitter', mapping them and comparing the result with the availability of the current data in BPS so that some suggestions for upgrading and advancing BPS in providing data to the public can be made. Some comments, especially about economics, from the public via Twitter were mined and classified by Decision tree method. Four categories data were produced: (1) Not in the field of economic, (2) in the field but not using BPS data, (3) in the field that may be using BPS data and (4) in the field that using BPS data. Tweets harvesting were done during two months period January 2016 til February 2016 by using Twitter Archiver. About 77.854 tweets that indicates to economic topic were collected. 2.360 data were used for training data. The result showed that 49.425 tweets were classified as economic topic. The majority of class was obtained from the tweets of economics but not using BPS data. 12 major keywords were obtained during this study. These keywords were generated using unigram, bigram, and trigram. In conclusion, we found that 3 primary keywords about economics were needed to be concerned by BPS: The economic policy package, Palm Oil, and Creative Economy.