Overview | Key Projects: Collaboration Continuum

Taxonomical Similarity Measures | NBER Full Text Extension | NBER Geolocation Project 

Data Overview

This dataset is designed to complement and enhance the traditional NBER patent dataset used in a great deal of management and international business research. It uses the original list of patents from the NBER dataset and extracts all data available on the USPTO website for each of these patents. This adds a range of useful data such as all classifications (not only primary class), search classifications, and un-truncated original text for fields like inventor locations.

The main contribution of this dataset is that it adds the full text of patent abstracts, all patent claims, and complete descriptions of patents. This is an incredibly rich dataset that can be used to bring a level of qualitative and quantitative rigor to patent data analysis not possible with currently available patent datasets. For example, patent classifications are one of the most commonly analyzed items of data in the NBER patent dataset. Each patent classification reflects the technology described in one of the patent's claims. However, each classification can only be assigned a single time to a patent (i.e., if 20 claims on a patent relate to a certain class/subclass and a last claim relates to a different class/subclass, the patent will be classified into these 2 class/subclass once each). Thus classification data is a limited and possibly biased representation of the actual unique technology represented by a patent and described in its claims.

The dataset is described more below:

NBER to Extension Match Percentages

Primary Patent Table

Title & Abstract Full Text Table

Claims Data Table

Description Full Text File

Please contact me if you would like to collaborate or have questions about any of my ongoing projects, datasets, or research capabilities.

robert.c.mcnamee at gmail