Interactive Sentiment Analysis of Social Media for Policy Decision Making

Foreign Affairs analysts monitor news media for evidence of shift in public sentiment on key political questions in countries with unstable regimes. Health Canada analysts monitor social media for public reactions to the outbreak of H1N1 flu and the government’s response. The analysts make recommendations to the government regarding public and defense policy decisions. Given the massive amounts of news and social media data, the analysts’ job is extremely labour intensive without any automated tools to help focus their attention to the documents that matter most. This project will focus on automating sentiment analysis with emphasis on its application to social media for the purpose of social policy related decision making. In contrast to a fully automated approach, the project will aim for a semi-automated approach to allow the analyst interaction with the sentiment analysis tools and tailor the output to his interests. The automatic sentiment analysis tools in the literature will be investigated for “hooks” that provide opportunities for user interaction. Finally, text visualization and human-computer interaction techniques will be investigated, to generate a complete prototype suitable for analyst use. The expected benefit for the partner organization, Mediabadger, is a dramatic improvement in the efficiency of social media monitoring, the main source of its revenue. Extension of the proposed system to consumer sentiment analysis will lead to a product addressing a very sizable commercial market.

Faculty Supervisor:

Dr. Vlado Keselj /Dr. Evangelos Milios /Dr. Kirstie Hawkey

Student:

Magdalena Jankowska

Partner:

MediaBadger

Discipline:

Computer science

Sector:

Media and communications

University:

Dalhousie University