Data-Driven News Generation for Indonesian Municipal Election

In order to fulfill the needs of journalistic automation, we develop automatic news generator that accepts structured data and user query and generates Indonesian news article. This paper employs template-based natural language generation in generating Indonesian municipal elections. Based on evaluation using Indonesian news characteristics as the evaluation metric by 15 linguistic experts and 25 active news readers, the average score of the generated news was 3.292 out of 4.