Philip Blair - Dealing With Names and Addresses Around the World | PyData Amsterdam 2024

Names and addresses vary dramatically across cultures and countries - assuming Western/English conventions leads to poor user experiences

Avoid using regex or fixed validation patterns for international names/addresses since they break when dealing with different formats and scripts

Key challenges with names include:

• Variable ordering (first/last name order differs by culture)
• Patronymics and matronymics
• Multiple last names
• Single names with no surname
• Different writing scripts and transliteration
• Marriage name changes varying by country

For address handling:

• Street number and name order varies by country
• Postal code formats differ significantly
• House number formats vary
• City/state conventions change

Best practices:

• Use existing libraries like PyPostal for address parsing
• Avoid unnecessary parsing - only structure data that’s truly needed
• Store original unstructured data alongside parsed versions
• Consider fuzzy matching instead of exact matching
• Use additional disambiguating data when available (birthdates, etc)

Commercial solutions are recommended for complex name matching needs like:

• Know Your Customer (KYC) verification
• Border screening
• Healthcare record linkage
• International sanctions checking

Focus on user experience by:

• Only collecting name/address details actually needed
• Supporting international formats
• Avoiding assumptions about name structures
• Making input forms flexible enough for global users