When doing quality rectification, fuzzy match based on city label (fuzzy) + dep code (strict) leads up to 98% of reidentification (with correct insee code) on birth places (base on real use case: french driving licence).
Same endpoint could still be used, with an option (like fuzzy=true or search=fuzzy) - using for example js-levenshtein library.
2 use cases :
when displaying some record to an enduser with "manual input historical data", UI suggest a rectification and an INSEE code for validation
when doing bulk quality rectification, massive call to this API could be useful (in this case, csv enrichment like https://adresse.data.gouv.fr/csv would be very useful)
When doing quality rectification, fuzzy match based on city label (fuzzy) + dep code (strict) leads up to 98% of reidentification (with correct insee code) on birth places (base on real use case: french driving licence).
Same endpoint could still be used, with an option (like fuzzy=true or search=fuzzy) - using for example js-levenshtein library.
2 use cases :