DATA

Download data

The patentCity database is available as an open access dataset (CC-BY-4).

Download patentCity database

Citation

If you use the data, please cite Bergeaud and Verluise (2021) and De Rassenfosse, Kozak and Seliger (2019)

BibTeX

@techreport{bergeaudVerluise2021,
  title={One Century of Innovation in Europe and the US},
  author={Bergeaud, Antonin and Verluise, Cyril},
  year={2021}
}
@article{deRassenfosse2019,
  title={Geocoding of worldwide patent data},
  author={De Rassenfosse, Ga{\'e}tan and Kozak, Jan and Seliger, Florian},
  journal={Scientific data},
  volume={6},
  number={1},
  pages={1--15},
  year={2019},
  publisher={Nature Publishing Group}
}

Chicago

Bergeaud, Antonin and Cyril Verluise. "One Century of Innovation in Europe and the US". 2021
De Rassenfosse, Gaetan, Jan Kozak, and Florian Seliger. "Geocoding of worldwide patent data." Scientific data 6, no. 1 (2019): 1-15.

Database schema

The database is available in 2 formats: csv and jsonl. Both contain the same data and are compatible with most SQL database engine and cloud data wharehouses such as BigQuery (GCP) and Athena (AWS) which should be your preferred way to work the patentcity full database.

jsonl or csv

The jsonl format allows nested data, hence is more compact. The csv format can be read off-the-shell by any library supporting tabular/structured data (e.g. pandas for python, dplyr for R, etc).

jsonl

Nested fields

• Dotted variables (e.g. patentee.is_inv) indicate nested fields.
• In the csv flavour of the data, the prefix (e.g. patentee.) is not reported. The database has already been flattened.

name	description	mode	type
publication_number	Publication number.	NULLABLE	STRING
publication_date	Publication date (yyyymmdd).	NULLABLE	INTEGER
country_code	Country code of the patent office.	NULLABLE	STRING
pubnum	Publication number.	NULLABLE	STRING
kind_code	Kind code.	NULLABLE	STRING
family_id	Family ID (DOCDB).	NULLABLE	STRING
cpc_code	Comma-separated list of cpc-codes.	NULLABLE	STRING
origin	Indicates the origin of the patentee data (PC: patentcity, WGP25: Worldwide Geocoding of Patent - slot 25, WGP45: Worldwide Geocoding of Patent - slot 45, EXP: expansion ).	NULLABLE	STRING
patentee.is_inv	True if the patentee is an inventor, else False.	NULLABLE	BOOLEAN
patentee.is_asg	True if the patentee is an assignee, else False.	NULLABLE	BOOLEAN
patentee.name_text	Name.	NULLABLE	STRING
patentee.person_id	Person ID (PATSTAT).	NULLABLE	INTEGER
patentee.name_start	Name start.	NULLABLE	INTEGER
patentee.name_end	Name end.	NULLABLE	INTEGER
patentee.occ_text	Occupation text.	NULLABLE	STRING
patentee.occ_start	Occupation start.	NULLABLE	INTEGER
patentee.occ_end	Occupation end.	NULLABLE	INTEGER
patentee.cit_text	Citizenship text.	NULLABLE	STRING
patentee.cit_code	Citizenship code.	NULLABLE	STRING
patentee.cit_start	Citizenship start.	NULLABLE	INTEGER
patentee.cit_end	Citizenship end.	NULLABLE	INTEGER
patentee.loc_text	Location text.	NULLABLE	STRING
patentee.loc_start	Location start.	NULLABLE	INTEGER
patentee.loc_end	Location end.	NULLABLE	INTEGER
patentee.loc_addressLines	Formatted address lines built out of the parsed address components.	NULLABLE	STRING
patentee.loc_locationLabel	Assembled address value for displaying purposes.	NULLABLE	STRING
patentee.loc_country	ISO 3166-alpha-3 country code.	NULLABLE	STRING
patentee.loc_state	First subdivision level(s) below the country. Where commonly used, this is a state code (for instance, CA for California).	NULLABLE	STRING
patentee.loc_county	Second subdivision level(s) below the country. Use of this field is optional if a second subdivision level is not available.	NULLABLE	STRING
patentee.loc_city	Locality of the address.	NULLABLE	STRING
patentee.loc_district	Subdivision level below the city. Use of this field is optional if a second subdivision level is not available.	NULLABLE	STRING
patentee.loc_subdistrict	Subdivision level below the district. Used only for India.	NULLABLE	STRING
patentee.loc_postalCode	Postal code.	NULLABLE	STRING
patentee.loc_street	Street name.	NULLABLE	STRING
patentee.loc_building	Building name.	NULLABLE	STRING
patentee.loc_houseNumber	House number.	NULLABLE	STRING
patentee.loc_longitude	Longitude.	NULLABLE	FLOAT
patentee.loc_latitude	Latitude.	NULLABLE	FLOAT
patentee.loc_relevance	Indicates the relevance of the results found; the higher the score the more relevant the alternative. The score is a normalized value between 0 and 1.	NULLABLE	FLOAT
patentee.loc_matchType	Quality of the location match. pointAddress: Location matches exactly as point address. interpolated: Location was interpolated.	NULLABLE	STRING
patentee.loc_matchCode	Code indicating how well the result matches the request. Enumeration [exact, ambiguous, upHierarchy, ambiguousUpHierarchy].	NULLABLE	STRING
patentee.loc_matchLevel	The most detailed address field that matched the input record.	NULLABLE	STRING
patentee.loc_matchQualityCountry	MatchQuality provides detailed information about the match quality of a result at attribute level. Match quality is a value between 0.0 and 1.0. 1.0 represents a 100% match. Here, matchQuality is defined at country level.	NULLABLE	FLOAT
patentee.loc_matchQualityState	Same at state level.	NULLABLE	FLOAT
patentee.loc_matchQualityCounty	Same at county level.	NULLABLE	FLOAT
patentee.loc_matchQualityCity	Same at city level.	NULLABLE	FLOAT
patentee.loc_matchQualityDistrict	Same at district level.	NULLABLE	FLOAT
patentee.loc_matchQualityPostalCode	Same at postalCode level.	NULLABLE	FLOAT
patentee.loc_matchQualityStreet	Same at street level.	NULLABLE	FLOAT
patentee.loc_matchQualityHouseNumber	Same at houseNumber level.	NULLABLE	FLOAT
patentee.loc_matchQualityBuilding	Same at building level.	NULLABLE	FLOAT
patentee.loc_key	Key used for statistical area mapping (internal use).	NULLABLE	STRING
patentee.loc_statisticalArea1	Name of the high level Statistical Area.	NULLABLE	STRING
patentee.loc_statisticalArea1Code	Code of the high level Statistical Area.	NULLABLE	STRING
patentee.loc_statisticalArea2	Name of the mid level Statistical Area.	NULLABLE	STRING
patentee.loc_statisticalArea2Code	Code of the mid level Statistical Area.	NULLABLE	STRING
patentee.loc_statisticalArea3	Name of the low level Statistical Area.	NULLABLE	STRING
patentee.loc_statisticalArea3Code	Code of the low level Statistical Area.	NULLABLE	STRING
patentee.loc_recId	Identifier of the input address in the response.	NULLABLE	STRING
patentee.loc_seqLength	Number of results for the corresponding input record.	NULLABLE	INTEGER
patentee.loc_seqNumber	Consecutively numbers the different results for the corresponding input record starting with 1.	NULLABLE	INTEGER
patentee.loc_source	Geocoding source (in [HERE, GMAPS, MANUAL]).	NULLABLE	STRING
patentee.is_duplicate	True if a patentee with the 'same' name has been detected in the same patent. Only one of the two is marked as duplicate.	NULLABLE	BOOLEAN
has_a	Whether the patent's family features an A kind code in the database [DOCDB family level].	NULLABLE	BOOLEAN
has_b	Whether the patent's family features a B kind code in the database [DOCDB family level].	NULLABLE	BOOLEAN
N	Number of patents in the family [DOCDB family level].	NULLABLE	INTEGER

csv

name	description	mode	type
publication_number	Publication number.	NULLABLE	STRING
publication_date	Publication date (yyyymmdd).	NULLABLE	INTEGER
country_code	Country code of the patent office.	NULLABLE	STRING
pubnum	Publication number.	NULLABLE	STRING
kind_code	Kind code.	NULLABLE	STRING
family_id	Family ID (DOCDB).	NULLABLE	STRING
cpc_code	Comma-separated list of cpc-codes.	NULLABLE	STRING
origin	Indicates the origin of the patentee data (PC: patentcity, WGP25: Worldwide Geocoding of Patent - slot 25, WGP45: Worldwide Geocoding of Patent - slot 45, EXP: expansion ).	NULLABLE	STRING
is_inv	True if the patentee is an inventor, else False.	NULLABLE	BOOLEAN
is_asg	True if the patentee is an assignee, else False.	NULLABLE	BOOLEAN
name_text	Name.	NULLABLE	STRING
person_id	Person ID (PATSTAT).	NULLABLE	INTEGER
name_start	Name start.	NULLABLE	INTEGER
name_end	Name end.	NULLABLE	INTEGER
occ_text	Occupation text.	NULLABLE	STRING
occ_start	Occupation start.	NULLABLE	INTEGER
occ_end	Occupation end.	NULLABLE	INTEGER
cit_text	Citizenship text.	NULLABLE	STRING
cit_code	Citizenship code.	NULLABLE	STRING
cit_start	Citizenship start.	NULLABLE	INTEGER
cit_end	Citizenship end.	NULLABLE	INTEGER
loc_text	Location text.	NULLABLE	STRING
loc_start	Location start.	NULLABLE	INTEGER
loc_end	Location end.	NULLABLE	INTEGER
loc_addressLines	Formatted address lines built out of the parsed address components.	NULLABLE	STRING
loc_locationLabel	Assembled address value for displaying purposes.	NULLABLE	STRING
loc_country	ISO 3166-alpha-3 country code.	NULLABLE	STRING
loc_state	First subdivision level(s) below the country. Where commonly used, this is a state code (for instance, CA for California).	NULLABLE	STRING
loc_county	Second subdivision level(s) below the country. Use of this field is optional if a second subdivision level is not available.	NULLABLE	STRING
loc_city	Locality of the address.	NULLABLE	STRING
loc_district	Subdivision level below the city. Use of this field is optional if a second subdivision level is not available.	NULLABLE	STRING
loc_subdistrict	Subdivision level below the district. Used only for India.	NULLABLE	STRING
loc_postalCode	Postal code.	NULLABLE	STRING
loc_street	Street name.	NULLABLE	STRING
loc_building	Building name.	NULLABLE	STRING
loc_houseNumber	House number.	NULLABLE	STRING
loc_longitude	Longitude.	NULLABLE	FLOAT
loc_latitude	Latitude.	NULLABLE	FLOAT
loc_relevance	Indicates the relevance of the results found; the higher the score the more relevant the alternative. The score is a normalized value between 0 and 1.	NULLABLE	FLOAT
loc_matchType	Quality of the location match. pointAddress: Location matches exactly as point address. interpolated: Location was interpolated.	NULLABLE	STRING
loc_matchCode	Code indicating how well the result matches the request. Enumeration [exact, ambiguous, upHierarchy, ambiguousUpHierarchy].	NULLABLE	STRING
loc_matchLevel	The most detailed address field that matched the input record.	NULLABLE	STRING
loc_matchQualityCountry	MatchQuality provides detailed information about the match quality of a result at attribute level. Match quality is a value between 0.0 and 1.0. 1.0 represents a 100% match. Here, matchQuality is defined at country level.	NULLABLE	FLOAT
loc_matchQualityState	Same at state level.	NULLABLE	FLOAT
loc_matchQualityCounty	Same at county level.	NULLABLE	FLOAT
loc_matchQualityCity	Same at city level.	NULLABLE	FLOAT
loc_matchQualityDistrict	Same at district level.	NULLABLE	FLOAT
loc_matchQualityPostalCode	Same at postalCode level.	NULLABLE	FLOAT
loc_matchQualityStreet	Same at street level.	NULLABLE	FLOAT
loc_matchQualityHouseNumber	Same at houseNumber level.	NULLABLE	FLOAT
loc_matchQualityBuilding	Same at building level.	NULLABLE	FLOAT
loc_key	Key used for statistical area mapping (internal use).	NULLABLE	STRING
loc_statisticalArea1	Name of the high level Statistical Area.	NULLABLE	STRING
loc_statisticalArea1Code	Code of the high level Statistical Area.	NULLABLE	STRING
loc_statisticalArea2	Name of the mid level Statistical Area.	NULLABLE	STRING
loc_statisticalArea2Code	Code of the mid level Statistical Area.	NULLABLE	STRING
loc_statisticalArea3	Name of the low level Statistical Area.	NULLABLE	STRING
loc_statisticalArea3Code	Code of the low level Statistical Area.	NULLABLE	STRING
loc_recId	Identifier of the input address in the response.	NULLABLE	STRING
loc_seqLength	Number of results for the corresponding input record.	NULLABLE	INTEGER
loc_seqNumber	Consecutively numbers the different results for the corresponding input record starting with 1.	NULLABLE	INTEGER
loc_source	Geocoding source (in [HERE, GMAPS, MANUAL]).	NULLABLE	STRING
is_duplicate	True if a patentee with the 'same' name has been detected in the same patent. Only one of the two is marked as duplicate.	NULLABLE	BOOLEAN
has_a	Whether the patent's family features an A kind code in the database [DOCDB family level].	NULLABLE	BOOLEAN
has_b	Whether the patent's family features a B kind code in the database [DOCDB family level].	NULLABLE	BOOLEAN
N	Number of patents in the family [DOCDB family level].	NULLABLE	INTEGER