/ClusterHN

Provides a way to quickly find posts of interest from Hacker News

Primary LanguagePython

ClusterHN

This application performs hierarchical agglomerative clustering based on the distances of cluster centroids. Information is obtained by scraping news links from HN.  Pearson correlation is used as the "distance" metric. A simple Tornado based server powers the application which updates is cache every 45 mintues with the most uptodate clusters.