The Patent Collaboration Network database is a dataset generated by pulling together data on all patents granted by US Patent and Trademark Office since 1975. The goal of the project is to represent the social network of inventors to study, for example, how certain patterns of collaboration influence creativity and innovation.

Patent data present a significant challenge in that they do not contain unique identifiers for every inventor. That is, there are many patents attributed to John Smith in the database, but these John Smiths are not all the same person. The first task in this project was to determine which inventors are which (a task known as author disambiguation) so that we can represent accurately the careers of individual inventors.

After individual authors are identified, we can construct a social network of inventor collaborations by drawing relationships between inventors who worked on the same patent. This collaboration network opens a number of new avenues of study, including which inventors are the most important or central in their field, how connectedness affects an inventor's productivity, which corporate structures are the most conducive to innovation, and whether legal changes impact substantially the flow of ideas among inventors. The dataset is thus of interest to scholars in a number of fields, from management science to history to political science.

A final goal of the project is to create an automated updating system so that the database can be as current as the weekly reports released by the USPTO on new granted patents. This enables scholars to track developments in the inventor community in real time.