Fuzzy Matching - Two data sources by address to bring back match record?

[KFarmer]

Hello -

I have a customer list and hospital roster list and I'd like to do a fuzzy match based on address (excluding name because naming conventions get wonky). Is it possible to do a fuzzy match based off address and then bring back the "Definitive ID" associated with each customer. Not likely a match for all accounts... I tried pulling something together but I'm not achieving what I'm looking to do. Any help would be appreciated!

Expected results would be the Customer list with the Definitive ID match.

[davidskaife · Accepted solution]

Hi @KFarmer 

How does this look? I've created a source column and unioned the two datasets together before fuzzy matching on 90% (see screenshot below with the setup), then join the matches to the Hospital Roster to obtain the Definitive ID, then join back to the Customer List.

It's a bit rough but hopefully gives you something to work with. Ideally you start with a really high % such as 96-98% and run through multiple fuzzy matches reducing the % as you go but i set it at 90% as that brought through your example above correctly.

Note there are duplicate matches appearing, which increase with each join, so look out for that.

Does that help?

[KFarmer]

Ah, that makes sense! Thanks so much!