Download inventors' patent data over the years

Here you can post your opinions, ask questions and share experiences on the PATSTAT product line. Please always indicate the PATSTAT edition (e.g. 2015 Autumn Edition) and the database (e.g. PATSTAT Online, MySQL, MS SQL Server, ...) you are using.
Post Reply

Lucatom3
Posts: 1
Joined: Wed Nov 02, 2022 6:40 pm

Download inventors' patent data over the years

Post by Lucatom3 » Wed Nov 02, 2022 6:52 pm

Hi, I'm Luca and I have a dataset consisting of firms.
I want to download data on the patents of the inventors of these firms during their entire careers (therefore also during periods when they worked for other companies).

At the moment, I am starting from the person_id of the firms to download the docdb_family_id. Once I have obtained the docdb_family_id, I input them to get the inventors of those firms' patents.

About the inventors, I obtain person_name, person_id and psn_id.
What is the best strategy to get all and only the patents of the inventors of my interest knowing that with the person_name there can be homonymy and the person_id can change over the years? Is the psn_id
reliable in this respect?

Thank you.
Luca


EPO / PATSTAT Support
Posts: 425
Joined: Thu Feb 22, 2007 5:33 pm
Contact:

Re: Download inventors' patent data over the years

Post by EPO / PATSTAT Support » Thu Nov 03, 2022 2:05 pm

Hello Luca,
There is no straigt forward method to identify inventor mobility in PATSTAT.
person_ids are fixed across PATSTAT releases, PSN_ID's can and will change, but the methodology (developped by KU-Leuven) stays the same.
Each person_id is linked to exactly 1 PSN_ID, and a PSN_ID can be linked to multiple person_id's.
So starting from 1 person_id you can via the psn_id see what more persons are "grouped".
Example: person_id = 2 (person_name: Lipponen, Markku; linked to applicant NOKIA) is linked to psn_id = 19669542 (in PATSTAT 2022a !) , which is it's turn linked to 13 "different persons". (a "person" is different when the data in the alternate key of tls206_person is unique)

Code: Select all

SELECT * FROM tls206_person
 where person_id= 2 or psn_id = 19669542
You can then see what companies are linked to those 13 persons via for example this query:

Code: Select all

SELECT distinct  inventor.person_id 
      , inventor.person_name 
      , inventor.person_name_orig_lg 
      , inventor.person_address 
      , inventor.person_ctry_code 
      , inventor.nuts 
      , inventor.nuts_level 
      , inventor.reg_code 
      , inventor.doc_std_name_id 
      , inventor.doc_std_name 
      , inventor.psn_id 
      , inventor.psn_name 
      , inventor.psn_level 
      , inventor.psn_sector 
      , inventor.han_id 
      , inventor.han_name 
      , inventor.han_harmonized 
	  ,a.appln_id
	  ,applicant.*
  FROM tls206_person inventor
  join tls207_pers_appln A on inventor.person_id = A.person_id and A.invt_seq_nr > 0
  join tls207_pers_appln B on A.appln_id = B.appln_id
  join tls206_person applicant on b.person_id = applicant.person_id and B.applt_seq_nr > 0 and applicant.psn_sector = 'company'
  where inventor.psn_id = 19669542
  order by a.appln_id
If you change the inventor.psn_id to 19669543, you will see a similar name linked to QUALCOMM Incorporated. You might conclude that the inventor worked for both companies (which also seems logical from a technical point of view), but looking at the patents itself, one can observe that patents were transferred from Nokia to QUALCOMM Incorporated. So even when the physical person is linked to multiple companies, it does not mean they were "working" there. You can find plenty of papers where inventor mobility is analysed using patent data. It might be worth to look into the methodologies.
https://scholar.google.com/scholar?hl=e ... stat&btnG=
PATSTAT Support Team
EPO - Vienna
patstat @ epo.org


Post Reply