Introduction -> give more focus to new methods and computational paradigms ?

[mpjoubertdebellefon]

Are innovation in the statistical processes so closely tied to the emergence of new sources ? As underlined in the Bucharest Memorandum that you quote, the amendment of the statistical process etc. is also due to the evolution of computational paradigms and tools. It seems to me that in our case at Insee, we are exploiting the contribution of new methods to process data that have already been processed for years (eg : NACE, PCS2020,...) For instance, FastText allows to better take into account unclear wording in textual descriptions, than what the previous deterministic model Sicore used to do. It is exactly how you problematise your article when you present the plan at the end of the introduction, yet since it is your main angle of analysis it might be interesting to underline this point also earlier in the introduction.

[avouacr]

I've removed a "new" to limit the emphasis at the very beginning of the intro. Otherwise, I'm using mostly "non-traditional data sources" which I think includes both big data sources as well as non-structured data from traditional sources (such as textual fields)