22799 Rapid and Longitudinal Analysis of H1N1 News and Social Media

Monday, April 19, 2010: 3:35 PM
Regency Ballroom VII
Alan Janssen, MSPH , Health Communication Specialist, Centers for Disease Control and Prevention

Background: Automated Immunization News for Managers (Auto-INFORM) system, sponsored by the CDC National Center for Immunization and Respiratory Diseases (NCIRD), captures and categorizes news articles for analysis of trends and issues in the areas of immunization and immunizable diseases.   AutoINFORM collects web feeds of daily news and business articles culled from thousands of worldwide newspapers, magazines, journals, etc.   The intelligence behind AutoINFORM’s linguistic processing allows for understanding and detection of precise media messages, facts, statements, and events that need to be monitored which would otherwise not be possible with competing keyword, categorization, or search based technologies.  Since April 15, 2009 AutoINFORM has collected 253,375 news reports; 57,128 blogs; and, 292,414 twitter comments on H1N1 influenza from a total universe of 12,522,201 reports.

Objectives: At the conclusuion of this presentation, the particpant should be able to describe three findings that emerged from H1N1 media analysis

Methods: Using the technology provided by Linguastat, Inc, AutoINFORM created customized reports that can identified, analyzed, and tracked specific H1N1 communication messages in news and social media.  H1N1 message reports included message counts & summaries, trend analysis, interactive discovery reports, and geospatial analysis and mapping views.  An unique Metascoring technology allows results to be quantified by specific metrics such as ratings information in order to estimate message reach

Results: The system idetified new communication messages/issues; measured media interest in select communication messages; provided daily geospatial analysis of news messages; provided communication reach estimates; and, produced a daily tag cloud to visually assess key communication elements along with roducing several longitudinal trend analyses.

Conclusions: The system was able to collect and analyze large volumes of news and social media reports and produce real time analysis of the content.

Previous Abstract | Next Abstract >>