Differences in the data between Patstat online and the EPO Statistics Centre

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

ChiniJ
Posts: 2
Joined: Wed May 18, 2022 5:58 pm

Differences in the data between Patstat online and the EPO Statistics Centre

Post by ChiniJ » Thu May 19, 2022 8:20 am

I wanted to inquire about differences in the data between Patstat online and the EPO Statistics Centre, when comparing the number of patent applications to European Patent Office (EPO), especially in the years 2019 and 2020:

Table: Comparison of overall number of patent application to EPO in Statistics Centre and PATSTAT
Year EPO StatCent PATSTAT difference
2012 148,562 149,085 -523
2013 148,027 153,378 -5,351
2014 152,703 156,996 -4,293
2015 160,004 155,455 4,549
2016 159,087 157,535 1,552
2017 166,594 160,714 5,880
2018 174,481 155,517 18,964
2019 181,532 68,243 113,289
2020 180,417 15,450 164,967

The query for Patstat:
SELECT DISTINCT a.appln_filing_year, a.appln_auth, count(distinct (a.appln_id))
FROM tls201_appln a
WHERE appln_filing_year BETWEEN 2012 AND 2020
AND appln_auth like 'EP'
GROUP BY a.appln_filing_year, a.appln_auth
ORDER BY a.appln_filing_year

Data from EPO Statistics centre: https://new.epo.org/en/statistics-centr ... lds=%5B%5D


We observe similar discrepancy also when comparing the number of patent application from various countries or within different technical fields.

Any ideas to what might be causing the differences especially for 2019 and 2020?


EPO / PATSTAT Support
Posts: 424
Joined: Thu Feb 22, 2007 5:33 pm
Contact:

Re: Differences in the data between Patstat online and the EPO Statistics Centre

Post by EPO / PATSTAT Support » Fri May 20, 2022 4:09 pm

Hello ChiniJ,
Statistics in the Patent Index and the STC are based on all recorded EP applications as filed during in the reporting period irrespective of their publication status.
Considering the publication delay, a substantial fraction of the reported applications are not yet published when the PI and STC are made available.
The non-published applications (also due to for example “withdrawal before publication”) do not appear in EPO online databases such as PATSTAT.
Furthermore, due to the early extraction date, some of the applications though filed in the reporting year, might not have been recorded yet. Therefore Patent Index statistics for the latest year are usually updated in the subsequent Patent Index (e.g. the figures for 2020 in the Patent Index 2020 were revised in the Patent Index 2021).
See also our clarification as to data sources: https://www.epo.org/about-us/annual-rep ... urces.html
PATSTAT Support Team
EPO - Vienna
patstat @ epo.org


ChiniJ
Posts: 2
Joined: Wed May 18, 2022 5:58 pm

Re: Differences in the data between Patstat online and the EPO Statistics Centre

Post by ChiniJ » Mon May 23, 2022 2:16 pm

Hi EPO / PATSTAT Support,

Thank you very much for your informative answer!
Can I also ask you whether the query we ran for counting the number of applications to EPO is correct?

Code: Select all

SELECT DISTINCT a.appln_filing_year, a.appln_auth, count(distinct (a.appln_id))
FROM tls201_appln a
WHERE appln_filing_year BETWEEN 2012 AND 2020
AND appln_auth like 'EP'
GROUP BY a.appln_filing_year, a.appln_auth
ORDER BY a.appln_filing_year
Furthermore, if we are interested in a number of applications from Belgian applicants to EPO, should we use this query?

Code: Select all

SELECT DISTINCT appl.appln_filing_year, appl.appln_auth, appl.appln_auth, person.person_ctry_code, count(distinct (appl.appln_id))
FROM tls201_appln as appl
JOIN dbo.tls207_pers_appln as pers_appln
on appl.appln_id = pers_appln.appln_id
JOIN dbo.tls206_person as person
on person.person_id = pers_appln.person_id
WHERE appln_filing_year BETWEEN 2012 AND 2020 
AND appln_auth like  'EP'
AND person_ctry_code like 'BE'
GROUP BY  appl.appln_filing_year, appl.appln_auth, person.person_ctry_code
ORDER BY appl.appln_filing_year


EPO / PATSTAT Support
Posts: 424
Joined: Thu Feb 22, 2007 5:33 pm
Contact:

Re: Differences in the data between Patstat online and the EPO Statistics Centre

Post by EPO / PATSTAT Support » Tue May 24, 2022 2:58 pm

Hello ChiniJ,
your first query is correct if you want to count all applications that have the intention to become patents with an EP effect. Just keep in mind that the EPO is also a PCT receiving office, and those patents will be found via the query below. Those PCT applications might or might not become EP applications when the PCT enters the regional phase.

Code: Select all

SELECT *
  FROM tls201_appln
  where receiving_office = 'EP' 
  and appln_filing_year BETWEEN 2012 AND 2020
Your second query is not 100% ok if you want to retrieve patens filed by BE APPLICANTS because your query will also include patents having a Belgian inventor (and for example a US applicant).
But was maybe what you intended to have anyway...

Code: Select all

SELECT DISTINCT appl.appln_filing_year, appl.appln_auth,person.person_ctry_code, count(distinct (appl.appln_id)) number_of_applications
FROM tls201_appln as appl
JOIN dbo.tls207_pers_appln as pers_appln
on appl.appln_id = pers_appln.appln_id
JOIN dbo.tls206_person as person
on person.person_id = pers_appln.person_id
WHERE appln_filing_year BETWEEN 2012 AND 2020 
AND appln_auth like  'EP'
AND person_ctry_code like 'BE'
AND applt_seq_nr > 0
GROUP BY  appl.appln_filing_year, appl.appln_auth, person.person_ctry_code
ORDER BY appl.appln_filing_year
PATSTAT Support Team
EPO - Vienna
patstat @ epo.org


Post Reply