Fuzzy string matching for addresses

I have some addresses that I need to match. What is the best way to match addresses that are written differently, but represent the same address?

For example:

  • 8/10 John Street and Unit 8 10 John Street

  • 19 Alexander Rd and 19 Alexander Road

  • 2/130 Mary St and 2/130-140 Mary St


1 Like


Perhaps you should use address standardizer / normalizer. You can find various tool or service for it on internet.


what do you want to match with?

@Pradeep_Shiv, each of the examples is a pair of strings that needs to be compared for similarity. Ideally, each of the pairs should be identified as matching since they represent the same address, i.e. “8/10 John Street” is the same as “Unit 8 10 John Street”

You can use edit distance algorithm , as per the percentage it retrieves you can proceed

I’m having a similar issue with address matching. Most of the time I end up doing a first word and postcode match - which matches the door number and post code, but except out any where the first word is not a number.

Would that not then give “Room 1, 19 Charlotte House, Newcastle-upon-Tyne, NE1 5XN” as a match to “Room 2, 19 Charlotte House, Newcastle-upon-Tyne, NE1 5XN” when they are infact different addresses?