Ask HN: An algorithm for an open source data mining application

If you start building an open source application for data mining, what algorithm would you like to see implemented first for data analysis? And why?