We intend to bring people together from diverse political affiliations to have meaningful conversation about issues and generate solutions. Patch is our first step in doing so.

Patch compiles news articles, social network comments, and user generated essays into clusters of relevant issues. It then performs feature extraction and analysis of the topics in those submissions. Using a fusion of standard Natural Language Processing techniques and the sophisticated APIs provided by Rhine, data such as information flow and subject similarity is extracted.

Some of the analysis Patch provides includes: Clustering of articles/issues with similar content Semantic correlation detection between words, sentences, or even articles. Frequency analysis of crucial Parts of Speech Synonym grouping and condensing into 'Lexical Lemmas'

Patch's polished web UI leverages many features of its powerful backend to present information to the user in a novel way. First, each article is analyzed to determine it's primary points. Next, the groups of sentences forming a point are assigned a representative sentence that most closely summarizes the point. The article as a whole is also assigned a sentence summarizing its collective points. The user can then read articles taking advantage of this added hierarchy, quickly extracting meaning and diving down into the details as desired.

The power of Patch isn't fully understood until you experience first-hand how closely it matches human intuition in determining the arguments of a piece of writing. So swing by our table tomorrow and say hello!

Share this project: