This week, I used AI to perform some address cleanup. Cities are notoriously known for being spelled incorrectly.
In the province of Quebec, things can get particularly interesting. For example, the following are actually all the same city:
- Saint Jerome
- Saint Jérôme
- St-Jérome
- St-Jérôme
I attached an .csv file to the prompt which had 2 columns: Count and MailingCity. Count is the number of times the version was used in a contact record.
I asked AI to create a 3rd column that mapped to the most used version of the city. So if the 1st version had a count of 10 and the rest were < 10, then set rows 2, 3, and 4 to the 1st spelling.
The results looked like
Count, MailingCity, MappedMailingCity 10, Saint Jerome, Saint Jerome 3, Saint Jérôme, Saint Jerome 8, St-Jérome, Saint Jerome 1, St-Jérôme, Saint Jerome
AI completed this task in seconds. I cannot imagine how long it would have taken me by any other method
The takeaway
Before AI, you couldn’t convince me to ever clean up city names. Now, bring it on!