I was analyzing a dataset that maps US Patent Classification (USPC) subclasses to Standard Industrial Classification (SIC) codes and product fields (description here, data available here, see zip file).

However, I found that the most common SIC class to which USPC subclasses were mapped was SIC code 21 (i.e. the US Postal Service!) Why is this? This seems surprising as I do not intuitvely regard the USPS as the most innovative industry in the US economy.

enter image description here

Some of my theories for why this is includes:

  • There may be peculiarities related to how government-owned patents are assigned SIC codes, since USPS is a government agency. This could inflate counts.
  • There may be subclasses related to generic business logistics/tracking that got assigned to USPS but are more broadly applicable.
  • The data is only looking at government-owned patents. So these are patents assigned to federal government agencies, not private companies. As a major government agency, it makes sense that the USPS would have a sizable patent portfolio.
  • Maybe I misunderstand the nature of this data, and it might just be looking at government-owned patents. Of US government agencies, the USPS might indeed file the most patents

1 Answer 1


They aren't.

This file does not at all represent individual issued patents but just shows a relationship between classifications that a patent might be assigned to by the USPTO and industries that are likely to make or use products that would fall into that classification.

Also, it has one step of indirection. If you look in the readme for the data set, the column you are looking at isn't the actual SIC code but a "sequence number" that they map on to an SIC code in the readme. Sequence number 1 stands for SIC code 20, for example.

To follow a specific segment of the file, take the patent class 16. It covers -


In the file there are about 500 entries for class 16's various subclasses. Most have a sequence number of 21. The readme shows 21 corresponds to

21 34- FABRICATED METAL PRODUCTS 34(except 3462,3463,348)

The two-digit 34 is Fabricated Metal Products Manufacturing

So we learn that most of the USPTO class for MISCELLANEOUS HARDWARE corresponds to the industry category of Fabricated Metal Products Manufacturing.

The SIC code with the most subclasses shown in your hostorgram is 43. According to the readme that coresponds to USPTO class 366.

43 366+ Electronic components and accessories and communications equipment 366-367

This is not an interesting data set and isn't really a data set at all, just a translation of one industrial coding system to another. And from your histogram we see tha Electronic components and Metal both Products Manufacturing maps on to many USPTO subclasses.

The histogram represents how fine-grained the respective classes are split into subclasses.

Irrelevant to the answer but out of curiosity I looked up SIC see image below, 21 is tobacco products.

table of first 22 SIC codes

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .