Our entire data processing pipeline is actually written in perl (that is several 10s of thousands of lines of code). So is the subcomponent that actually pipes the data to Geckoboard.