Development of Application Providing Public’s Perspectives on Official Statistical Indicators

Abstract

BPS-Statistics Indonesia, as an official data producer, puts data quality as a top priority. Public acceptance and trust in data reflect data reliability which is one of the data quality indicators. The existing survey that collects user acceptance, trust, and perspective to data produced by BPS can only reach data users in a limited number. Perspectives from a large number of data users cannot be captured. This research aims to build an application that collects and provides users’ perspectives and sentiment to official statistics sourced from online news. One feature in this application is Named Entity Recognition, which extracts public perspectives in entities such as names, organizations, statistical indicators, quotes or opinions, etc. This application objectively measures the sentiment of news discussing or citing statistical indicators. This application also facilitates BPS to do social network analysis to understand the relationships between fellow data users for each statistical indicator. The implicit goal is to effectively provide insights into how frequently society uses and refers to statistical indicators produced by BPS in any domain and their perspective on data. All models and features provided in this application have been evaluated based on standard performance metrics.