Bridging the Publication Window Blackout

Patent filings are a huge source of data, but there is a fundamental flaw.  Unpublished applications make insights stale.  Data science was used to bridge that gap in a patent trends study to give timely predictions for different technologies in a 2018 patent study co-authored by the presenter.  A 2018 dataset of art unit filings was used to pierce through the non-publication window to give timely trends in a dozen broad categories with many more sub-categories.

Layering in machine learning allows grander ambitions.  Inventors want to know how unique their innovation is before patent filing.  Patent trends augmented with multiple sources of information can be used to instantly give a current indication of patent popularity.  Knowing in advance can accelerate effort toward filling, innovating and product development.