When fed a stream of data, this applies a filter to the stream to extract items that are interesting, and reports the frequency that those items appear within the stream.
Eventually, this will be something nice and efficient, but for now a python version will be in /prototype