Difference between revisions of "UVM Patent Filings"

From edegan.com
Jump to navigation Jump to search
Line 4: Line 4:
  
 
Accordingly, I retrieved all of UVM's patent filings to the USPTO from [https://cloud.google.com/blog/products/gcp/google-patents-public-datasets-connecting-public-paid-and-private-patent-data Google Patents Public Dataset] using the following SQL in BigQuery:
 
Accordingly, I retrieved all of UVM's patent filings to the USPTO from [https://cloud.google.com/blog/products/gcp/google-patents-public-datasets-connecting-public-paid-and-private-patent-data Google Patents Public Dataset] using the following SQL in BigQuery:
  WITH Patent_Matches AS
+
  WITH Patent_Matches AS (
  (SELECT PARSE_DATE('%Y', SAFE_CAST(ANY_VALUE(patentsdb.filing_date) AS STRING)) AS Patent_Filing_year,
+
  SELECT  
        patentsdb.application_number AS Patent_Application_Number, ANY_VALUE(assignee_name.name) AS AssigneeName
+
PARSE_DATE('%Y%m%d', SAFE_CAST(ANY_VALUE(patentsdb.filing_date) AS STRING)) AS Patent_Filing_Date,
FROM `patents-public-data.patents.publications` AS patentsdb, UNNEST(assignee_harmonized) AS assignee_name
+
patentsdb.application_number AS Patent_Application_Number,
WHERE LOWER(assignee_name.name) LIKE '%univ vermont%' AND patentsdb.country_code = 'US'
+
ANY_VALUE(assignee_name.name) AS AssigneeName
GROUP BY Patent_Application_Number
+
FROM `patents-public-data.patents.publications` AS patentsdb,
  )
+
UNNEST(assignee_harmonized) AS assignee_name
  SELECT Patent_Filing_year, COUNT(Patent_Matches.Patent_Application_Number) AS Number_of_Patent_Applications
+
WHERE LOWER(assignee_name.name) LIKE '%univ vermont%' AND patentsdb.country_code = 'US'
  FROM Patent_Matches  
+
GROUP BY Patent_Application_Number
  GROUP BY Patent_Filing_year
+
),
  ORDER BY Patent_Filing_year;
+
Date_Series_Table AS (
 +
SELECT day, 0 AS Number_of_Patents
 +
FROM UNNEST (GENERATE_DATE_ARRAY(
 +
  (SELECT MIN(Patent_Filing_Date) FROM Patent_Matches),
 +
(SELECT MAX(Patent_Filing_Date) FROM Patent_Matches)
 +
)) AS day
 +
)
 +
  SELECT SAFE_CAST(FORMAT_DATE('%Y-%m',Date_Series_Table.day) AS STRING) AS Patent_Date_YearMonth, COUNT(Patent_Matches.Patent_Application_Number) AS Number_of_Patent_Applications
 +
  FROM Patent_Matches
 +
RIGHT JOIN Date_Series_Table
 +
ON Patent_Matches.Patent_Filing_Date = Date_Series_Table.day
 +
  GROUP BY Patent_Date_YearMonth
 +
  ORDER BY Patent_Date_YearMonth;
  
 
Note that UVM's harmonized assignee name is UNIV VERMONT. Also, I could only go up to 2017, as patent applications are typically not disclosed for 18 months to protect their investors.
 
Note that UVM's harmonized assignee name is UNIV VERMONT. Also, I could only go up to 2017, as patent applications are typically not disclosed for 18 months to protect their investors.

Revision as of 18:41, 17 February 2020

UVM has recently experienced a dramatic and precipitous drop in rankings of patents granted to universities. However, those rankings use ten-year cumulative patent grants, and patents can take several years to be processed by the USPTO. Moreover, UVM could conceivably have been granted more patents but failed either to keep up with its peers or with patent inflation.

BigQuery Code

Accordingly, I retrieved all of UVM's patent filings to the USPTO from Google Patents Public Dataset using the following SQL in BigQuery:

WITH Patent_Matches AS (
	SELECT 
	PARSE_DATE('%Y%m%d', SAFE_CAST(ANY_VALUE(patentsdb.filing_date) AS STRING)) AS Patent_Filing_Date,
	patentsdb.application_number AS Patent_Application_Number,
	ANY_VALUE(assignee_name.name) AS AssigneeName
	FROM `patents-public-data.patents.publications` AS patentsdb,
	UNNEST(assignee_harmonized) AS assignee_name
	WHERE LOWER(assignee_name.name) LIKE '%univ vermont%' AND patentsdb.country_code = 'US'
	GROUP BY Patent_Application_Number
),
Date_Series_Table AS (
	SELECT day, 0 AS Number_of_Patents
	FROM UNNEST (GENERATE_DATE_ARRAY(
	(SELECT MIN(Patent_Filing_Date) FROM Patent_Matches),
	(SELECT MAX(Patent_Filing_Date) FROM Patent_Matches)
	)) AS day
)
SELECT SAFE_CAST(FORMAT_DATE('%Y-%m',Date_Series_Table.day) AS STRING) AS Patent_Date_YearMonth, COUNT(Patent_Matches.Patent_Application_Number) AS Number_of_Patent_Applications
FROM Patent_Matches
RIGHT JOIN Date_Series_Table
ON Patent_Matches.Patent_Filing_Date = Date_Series_Table.day
GROUP BY Patent_Date_YearMonth
ORDER BY Patent_Date_YearMonth;

Note that UVM's harmonized assignee name is UNIV VERMONT. Also, I could only go up to 2017, as patent applications are typically not disclosed for 18 months to protect their investors.

Results

UVMPatentFilings.png

UVM had its best year on record in 2013 with 24 filings. However, it made just 4 filings in 2015, which was its lowest count since 1998, and then just a single filing in both of 2016 and 2017, which tie 33-year lows. Of course, UVM's own data may show material for 2018 and 2019. But, as it stands, the results are suggestive of a structural change in UVM's patenting activities somewhere around 2013.


Notes

UVM's 26th President, Tom Sullivan, began his term in July 2012, and hired much of his administration in the following two years. Suresh Garimella became the 17th President of UVM in October 2019. Garimella previously had notable success in spurring the development and commercialization of technology at Perdue University, where he was the inaugural Executive Vice President for Research and Partnerships.