Evaluating the influence of the reporting of natural disasters on public understanding of their causes and possible prevention solutions

This project will investigate correlations between natural disasters, climate change, and the reporting and understanding of such events. It will deploy a Natural Language Processing (NLP) framework for text analysis, used to understand the content of news articles and messages on social media, and methods from journalism and machine learning strategy to examine the ways in which news stories are reported which generate emotions or sentiments polarity by using the language of risk, fear and anxiety.

Examination of the style and content of news stories will establish the degree to which framing of information (amplifying, diminishing or distorting the event) has been undertaken. An open-source NLP software, GATE, will be applied in this project for analysing the news articles from different types of sources, and investigating the use of language in the articles. The latest word-embedding will also be applied for analysing the similarity between words, sentences and documents. News articles will be classified according to sentiment and emotion, to understand what attitudes and motivations appear in different places and time periods.

GATE team (including Ye Jiang) wins first prize in the Hyperpartisan News Detection Challenge

Ye Jiang’s homepage


Dr Diana Maynard

Department of Computer Science


Professor Shaun Quegan

School of Mathematics and Statistics