Zoran Obradovic - NSF Assessing Influence of News Articles on Emerging Events

Home       Bio/CV       Research       Publications       Teaching       Lab       Heritage       Contact

EAGER:IIS/III: Assessing Influence of News Articles on Emerging Events

Principal Investigator:

Zoran Obradovic, L.H.Carnell Professor of Data Analytics, Temple University

Co-PI:

Eduard Dragut, Assistant Professor, Computer and Information Sciences, Temple University

Abstract

Social media and news articles play an important role in documenting daily societal events. News outlets host social media platforms that facilitate users to engage in debating daily news topics. For example, the social networks at NY Times, The Guardian, and Washington Post have more than 130K users each. Together, they constitute a considerable segment of the varied opinions of the society at large. The difficult and high risk problem addressed in this project is that of transforming the streams of social media chatter at hundreds of news outlets into data signals from which to mine those signals foretelling the imminence of an (important) event, and to develop sound predictive analytics on top of those signals. The project benefits multiple segments of society, such as social scientists and policy makers, because the results of the proposed project provide tools to predict important real-life events using indicators observed on social media. There is growing interest in mining social media streams for early detection of (important) events, like crisis detection (and response) and predicting social unrest.

The objective of this project is to assess the feasibility of leveraging the trend of past social response to news articles observed over a few hundred social media streams to detect the emergence of social, economic, and political events. This project seeks creating a proof of concept that works with a few hundred social communities from news outlets. Specific aims consist of (i) developing methods for automatic data collection and (ii) efficient predictive modeling at that scale. The results (e.g., software tools) are made available to benefit researchers in academia and industry. Free, open-source software for implementing the developed techniques is distributed to enhance existing research infrastructure. The educational component of the project includes the involvement of graduate and undergraduate students' training and research and the incorporation of research projects/results in appropriate courses.

© 2001-2013 Center for Data Analytics and Biomedical Informatics, Temple University