Quick and Dirty Address Matching with LibPostal
Most businesses have databases of previous customers, and data analysts will frequently be asked to join arbitrary data to the customer tables in order to provide analysis.
Unfortunately joining address data together is notoriously difficult:
- The same address can be expressed in many ways
- The parts of addresses are not always clear
- There are valid lexically very similar addresses very nearby any given address