I've been working on a new method for analyzing and parsing datasets to identify and isolate subgroups of a population without foreknowledge of any subgroup's characteristics. While the method works well enough with artificial data samples (i.e. datasets created specifically for the purpose of identifying and segregating subsets of the population), I'd like to try testing it with live data.
What I'm looking for is a freely available (i.e. non-confidential, non-proprietary) data source. Preferably one containing bimodal or multimodal distributions or being obviously comprised of multiple subsets that cannot be easily pulled apart via traditional means. Where would I go to find such information?
Best Answer
Also see the UCI machine learning Data Repository.
http://archive.ics.uci.edu/ml/