Purpose The purpose of the 2012 Presidential Election demo is to analyze the difference between what people say on Twitter and what news articles say about the presidential candidates, Barack Obama and Mitt Romney. This demo takes a look at the sentiment towards the candidates during the months leading up to the election based on people’s tweets and the news articles that reference them. Tag words mentioned in tweets and news articles were also highlighted and explored in this demo as well as the categorization of the content as it relates to top topics discussed during the election.
Data The data used in this demo comes from two main sources: Twitter and online news articles. Industrial CodeBox’s (ICB) Twitter connector was used to collect the tweets for the presidential candidates. ICB also created the connector that was used to interact with Saplo’s API. Saplo’s tools provided the sentiment analysis for tweets and news articles, the selection of tag words in the content as well as the categorization of the tweets and news articles by top election topics. Below is a list of the news articles that Saplo crawled for this demo. A variety of news sources were selected in attempt to get a broad range of views across regions and political parties. News site CNN Fox News The Los Angeles Times The Guardian The New York Times The Wall Street Journal (online version) The Telegraph USA Today BBC Gulf News The Local Sweden The Local Germany Yle Huffington Post Huffington Post The Washington Post Miami Herald Charlotte Observer The Kansas City Star Cleveland Plain Dealer Chicago Tribune Star Tribune 2012 Presidential Election Demo
Analysis Below is some of the analysis that can be performed in the 2012 Presidential Election demo by sheet: Dashboard – This sheet highlights the overall sentiment for the presidential candidates and provided a breakdown of the sentiment specifically for tweets and articles. Sentiment for the candidates is also provided in the months leading up to the election in a line chart so the changes in sentiment over time can be viewed. The number of tweets and articles analyzed is provided as well as the top tags found for each candidate. The Twitter Button extension is also used on this sheet providing an easy way to tweet the demo to others. Sentiment Cloud – This sheet makes use of a word cloud extension that allows us to see which tag words were used the most for each candidate as well as the sentiment related to that tag word. The larger the tag word, the more often the tag word was mentioned for the respective candidate. The smaller the tag word, the less frequently the tag word was used. The color of the tag word corresponds to the sentiment. With regards to color, blue is used for Barack Obama and red is used for Mitt Romney. The darker the tag word, the more positive the sentiment. The lighter the tag word, the more negative the sentiment. Note that the word clouds used on this sheet excludes the words Obama, Barack Obama, Romney and Mitt Romney since they are used so frequently. The Sentiment Cloud sheet also examines the number of tweets and articles there were by topic and candidate to see which topics were referred to when discussing the candidates. The number of tweets and articles created/published over time is also explored. Positive Sentiment – This sheet looks only at positive sentiment for the candidates over time and how they differ for each candidate with regards to topic and source (Twitter or news article). Topic Ranking – A stream chart extension is used on this sheet to illustrate the popularity of election topics over time. The number of tweets and articles regarding the topic when discussing Barack Obama and Mitt Romney can be viewed as well as how the ranking of the topics change over time. Tweets vs Articles – This sheet also used the word cloud extension. On this sheet, the number of tags for tweets and articles can be compared for both candidates. The larger the tag word, the more often the tag word was mentioned and the smaller the tag word, the less frequently the tag word was mentioned. The color of the tag word indicates which candidate the tag word was used with the most. If the tag word is blue, then it was used more often when Barack Obama was discussed. It the tag word is red, it was used more often when Mitt Romney was discussed. If the tag word is purple then it was used equally when Obama and Romney were discussed. This comparison between tags in tweets and
2012 Presidential Election Demo
articles can be explored. Note that the word clouds used on this sheet excludes the words Obama, Barack Obama, Romney and Mitt Romney since they are used so frequently. About – The About sheet provided information about ICB and Saplo and their offerings as well as the extensions used in the demo.
Conclusion The 2012 Presidential Election demo explores the difference between our voice (Twitter) and the voice of the news and how they feel about the presidential candidates Barack Obama and Mitt Romney. The demo provides an interesting perspective on the presidential candidates from the sentiment when they are mentioned in tweets and news articles to the tag words and topics that are found when they are referenced in tweets and news articles.