Geocoding of worldwide patent data.

Area of Science:

• Intellectual Property Analysis
• Geospatial Data Science
• Innovation Studies

Background:

• Traditional patent analysis often overlooks geographic distribution.
• Established patent offices may create biases in data.
• Understanding inventor and applicant locations is crucial for innovation mapping.

Purpose of the Study:

• To create a geocoded database of patenting activity.
• To enable fine-grained geographic analysis of innovation.
• To mitigate biases associated with traditional patent data sources.

Main Methods:

• Geocoding of inventor and applicant locations from 18.8 million patent documents.
• Allocation of geocoded data to countries, regions, and cities.
• Imputation of missing address information using subsequent patent family filings.

Main Results:

• A comprehensive dataset with geographic coordinates for over 18.8 million patent documents.
• Geocoded data spanning over 30 years, allocated to specific geographic levels.
• Methodology to address missing address information, enhancing data completeness.

Conclusions:

• The developed database facilitates detailed geographic studies of patenting activity.
• This resource can reveal innovation patterns at a granular level.
• The approach reduces bias, offering a more equitable view of global innovation.