[P] Better ways to clean lots of text?

I'm working my way through 10k addresses I scraped from my county's eviction dockets. The goal is to geocode them so I can map them by neighborhood.

They're a complete mess. I'm fixing things, but it's so very slow. It's a long iterative process where I find a problem/pattern of problems, write a regex pattern to catch and fix it, then iterate. (ex: I have a dictionary to look up spelled out street name suffixes, and I just found someone …

machinelearning text