The treemap graphic shows the top 575 licenses in the DPLA corpus, divided into four clusters: those in copyright, those with no copyright restrictions, Creative Commons records and those where copyright is unknown, representing licenses that had at least 200 records associated with them. This was a somewhat arbitrary decision, but showing many more significantly slows down the loading and speed of dynamic features of the graphic. Other records were removed if the rights statement from the record didn't seem to make sense.
Blocks of the same color denote related records. Hover over a block to see the full license and how many records are associated with it. Click a box to expand the treemap to only show records of a similar license type. Click again to go back to the full list.
Upgrades to the visualization might include showing the full record to the right of the treemap instead of on top of the individual records. Another thought is linking out to a search of DPLA records with that particular license. My methodology was admittedly naive, and wouldn't stand up as a work of data science, but I think it gives a solid flavor of what's in the DPLA. Click here for a short overview of how I set the data up.
Note: Treemap visualization heavily based on http://mbostock.github.io/d3/talk/20111018/treemap.html