At the music publisher where I work, every quarter we have something called an Accounting Week, where we take statements from vendors (streaming services i.e. Spotify, sub-publishers) to import all transactions that took place in the particular quarter and subsequently pay songwriters. Part of the problem is that many times songs titles come with slight punctuation variations that mess with automatic text matching, so we are forced to manually select the right song with a dropdown menu when the time comes.
I read that through graph matching (from some branch of AI) one is able to rapidly match all song name variations together (along with their deriving songwriter name variations). You can imagine these as nodes. A graph is supposed to be helpful because it represents a relationship of variables; otherwise one would individually have to look up variable names with a for loop or the like. I am leaving the relevant article here: https://medium.com/blokur/match-that-consolidating-musical-works-registrations-with-graph-matching-d0e7bfe47303
Thank you in advance!