Detecting News Event from a Citizen Journalism Website Using Tags

摘要

The accelerated news cycle and constantly emerging news-worthy events have led to 'citizen journalism' where people who are non-journalists collect, analyze and disseminate news pieces. This paper seeks to leverage tags drawn from iReport, an active citizen journalism Website to detect major news events. The goal is to examine the coverage and efficacy of news detected in iReport vis-a-vis those reported in the mainstream media. The data collection procedure involved manually culling major news events reported in Fox News between April 8 2008 and June 6 2008. Additionally, 81,815 tags from 15,216 documents were drawn from iReport during the same study period. Relative frequencies of all unique tags were used to check for spikes and bursts in the dataset. The results show that out of the 10 major news events reported in Fox News, five could be detected in iReport. This paper concludes by presenting the main findings, limitations and suggestions for future research.