Research into invention, innovation policy, and technology strategy can greatly benefit from an accurate understanding of inventor careers. The Fung Institute generates and distributes a variety of software useful for innovation research based on patent data.
Large-scale studies of inventors are challenging, as the United States Patent and Trademark Office does not provide unique inventor identifiers. Many scholars of invention have implemented ad-hoc disambiguation methods based on string similarity thresholds and string comparison matching, but such methods have been shown to be vulnerable to a number of problems that can adversely affect research results. The disambiguation method used by the Fung Institute Inventor Database is represented below and based on the working paper by R. Lai, A. D’Amour, A. Yu, Y. Sun, V. Torvik, and L. Fleming, “Disambiguation and Co-Authorship Networks of the U.S. Patent Inventor Database.” You can view the complete input and data results from the original disambiguation, the revised data, and the original and revised code.
To promote transparency and reproducibility, the Fung Institute provides software source code and the raw data used for research purposes via the Github social coding site. The data and tools are free to download and use.