Data Hacks is a new library developed at popular URL-shortener bit.ly that consists of a set of command line tools to assist in quick and dirty data analysis. For instance, the library has tools to calculate 95 percentile values, a histogram display, sample to a % of stdin, and a tool to pass stdin to stdout for a set time period, all controlled from the command line. Such tools could be used to grab a sample from a log file of a running system. When combined with other command line tools, this could be useful for a real-time evaluation of the effect of configuration changes.
As a bit of coincidence, they also ended up with horizontal bar graphs made in ASCII [jehiah.cz].
You can download it here.