The SentiWords Tweet system generates real-time, visual representations of public opinion mined from topic related tweets using sentiment analysis tools.
As the national broadcaster, RTÉ report news as it unfolds to the nation. Described as the first social media election in Irish history, RTÉ collaborated with ADAPT in order to report on insights derived from social media during the General Election 2016.
Addressing the challenge, the ADAPT team developed a system that uses a sophisticated combination of keywords and common election-related hashtags to record the volume of tweets about candidates, parties and issues, and measure the sentiment expressed in them. The interface presents a graphical analysis of public opinion allowing the user quickly gain insights and measure reaction to events.
The ADAPT Solution
‘SentiWords Tweet’ is a responsive web interface that links to a set of underlying harvesting scripts that provide automatic, real-time generation of the quantity and sentiment associated with topic specific tweets. Data collected by the system can be easily filtered by criteria including candidate, party, date and time. The graphical dashboard provides easy recognition of trends and insights into public opinion.
The tool provides automatic sentiment classification of tweets using opinion lexicons that associate sentiment polarity for words. Sentiment is scored positive, negative or objective. Linking Twitter data to metrics of political support not only provides news agencies with access to public opinion as it forms but could also provide a viable replacement or supplement to traditional survey-based techniques with social media based strategies.
"Having the ability to access and pull data at a time of our choosing was a great help to us when reporting during the General Election." Will Goodbody, Science and Technology Reporter, RTÉ.