We introduce mldata.org, an on-going effort to build a data and benchmark repository for the machine learning community (http://mldata.org/about/motivation/). Compared with existing data repositories such as the UCI machine learning repository, the aims of mldata.org are to provide

(a) a dynamic site which is frequently updated by a self-moderating user community

(b) machine learning benchmarks and challenges in addition to a store of datasets

(c) an easy-to-use API to facilitate and automate interaction with the repository from popular scientific computing environments (such as python, matlab, and R)