Class Projects - Fall 2013

OpShop

OpShop analyzes news stories and presents readers with quotes from political leaders that relate to the story.

Screencast for OpShop, a proof of concept technology that analyzes news stories and presents readers with quotes from political leaders that relate to the story.
What it does
OpShop is a web application that can take any news article and find relevant quotes for readers from political leaders, giving readers a wide yet concise view of the political discourse surrounding a given issue. Trying to find a full range of opinions about a given topic can be a difficult process: The sheer number of news sources is overwhelming, the nature of most news coverage tends to be politically polarized and a simple web search often comes up with ambiguous, imprecise results. OpShop gives users a sampling of congressional opinions surrounding the topic in a particular news article. Quotes are retrieved from other articles and displayed side by side
How it works
To use OpShop, users enter the URL of a news article in the OpShop site.  The application takes the article link and uses Alchemy API to extract relevant keywords from the article. OpShop then accesses a congressional database containing the name, party, and partisan ranking (based on rating systems obtained from Americans for Democratic Action and the American Conservative Union) to identify members from different parts of the political spectrum. The system then randomly selects the names of four members of Congress, making sure the selected members lie in different categories along the political spectrum. Finally, OpShop creates four queries built using the Congress member names and the extracted keywords, which it runs through the Bing News API to obtain three articles per congressperson, yielding a total of 12 article.

Once the Bing News API returns these articles, OpShop uses Alchemy API once more to extract quotes by the selected congresspeople from the articles. For each of the four names, if a quote is found, the application saves both the quote and related article. When OpShop has found quotes and articles for all four congresspeople, it returns to the user the information (name, party, state) of the Congress members, their extracted quotes and links to the selected articles.
Next Steps
  • OpShop could be made into a browser extension, rather than a standalone web application.
  • Find or design a more powerful language-processing tool that would increase the speed of text extraction and the relevance of keywords and quotes extracted.
  • Extend database to include public figures outside of Congress, such as media pundits and other government officials. Public figures could also be chosen in a more targeted way, taking into account geographical location, areas of expertise and other such factors.
  • Develop dimensions beyond Liberal–Conservative to “rank” our public figures and take into account specific areas of policy, such as the economy or the environment.

Connect

Student Team: Kelly Gonsalves, Salil Gupta, Alex Sanz

Faculty Guidance: Larry Birnbaum, Rich Gordon