For my job, we're working on car parking data, looking at snapshots of different days for different locations. Some days are 30% full, some days are 50% full, and so on. We have 84 days and 30 locations. We get quite a variety of busy-ness for each location. A simple median and standard deviation isn't enough. I need something more expansive. I'm wondering what classification methods there are?

Basically, if you've got a bunch of box plots of different shapes, what is a formal way of classifying them?