Disease surveillance in Indonesia through Twitter posts

Abstract

Social media data has become popular resources for various research topic such as public health. One of the popular research directions is to use social media data to detect if there is an epidemic disease emerging in a certain area. This paper presents a framework for mapping the emergence of disease in Indonesia using data from Twitter. The framework is built upon several methods which consist of classification using SVM, clustering using K-Means, and a named-entity recognizer to extract location names. Our research successfully identifies tweets indicating disease emergence and generates a relatively accurate map visualization. Thus, we believe that using Twitter may help Indonesia government officials to get an overview of the spread of disease in Indonesia in a relatively short time.