Title: Diagnosing editorial strategies of Chilean media on Twitter using an automatic news classifier

Abstract: In Chile, does not exist an independent entity that publishes quantitative or qualitative surveys to understand the traditional media environment and its adaptation on the Social Web. Nowadays, Chilean newsreaders are increasingly using social web platforms as their primary source of information, among which Twitter plays a central role. Historical media and pure players are developing different strategies to increase their audience and influence on this platform. In this article, we propose a methodology based on data mining techniques to provide a first level of analysis of the new Chilean media environment. We use a crawling technique to mine news streams of 37 different Chilean media actively presents on Twitter and propose several indicators to compare them. We analyze their volumes of production, their potential audience, and using NLP techniques, we explore the content of their production: their editorial line and their geographic coverage.