There's an issue with data concordance (I think that's the term) in that, for example, Snekma is listed as having a few RU patents but that's actually the same company as SNECMA (French aeronautical company) which is listed as having many more patents; but that company is now known as Safran.

A proper historical analysis of companies would need to account for mergers/splits/renames of companies themselves - data which is only really able to be inferred (at best) from patent sources. You'll need company merger records etc..

I wonder if they're also handling misspellings in company data? (Eg Bpsch instead of Bosch, this sort of thing does sometimes get printed on patent documents).

Chinese company names are one of the harder cases, I feel, for example Edwards (pump manufacturer) presence in China appears to go by "Aidehua Vacuum" (which seems to be a roman-script Chinese transliteration of Edwards Vacuum).

It's a hard problem to attend but I think you're going to miss a lot of detail in company-focused analysis if you don't tackle it.

This is all personal opinion and in no way relates to my work."