Okay, so I'm really excited for this particular update. One of the challenges in reviewing patents / patent applications, in my view, has been understanding the context of the patent.
That is, answering questions like: what else is out there, who's working on similar problems, how new is this idea / concept, and is this trending up or down?
Well, I've got an initial version of what I'm calling the "Instant Patent Landscape" up & running, which goes some way towards answering these questions. These reports are generated by finding a set of similar patents to the initial document, and performing analysis on the result set.
Here are some examples - all of the analytics you'll see on each page is generated on demand.
For those curious:
Back-end: PostgreSQL, Python, NLTK/SpaCy, plot.ly, Django