Automatic Newsletter Generation

In the age of information inflation, news is no longer produced or consumed in a centralized fashion. The media landscape has changed radically, mainly because of: i) the instantaneous rate at which individuals publish news-worthy content, ii) the vast reachability of this content by broad audiences, and iii) the lack of regulation and quality control. The modern news landscape consists of mainstream news outlets that are supported, complemented, and often criticized by independent or alternative media channels. Although media companies are typically responsible for discovering and communicating news to the people, information paths are becoming increasingly convoluted with social networks acting as a diffusion medium.

NewsTeller is a real-time news analytics platform that provides a wide variety of tools for mapping the media landscape, monitoring the reach and the stance of news consumers, and providing quality indicators for millions of news articles and thousands of sources. In this project, we will focus on investigating techniques for automatic newsletter generation. These techniques will summarize the important news events given a certain time frame and topic while being provenance-aware by crediting the original sources. In the context of this project, we will utilize existing news article features (e.g., social media popularity), and develop models for extracting new ones (e.g., geolocation). The methodology will be integrated and made immediately available on the platform.