slu-openGIS / postmastr

R package for Processing and Parsing Untidy Street Addresses
https://slu-opengis.github.io/postmastr/
GNU General Public License v3.0
37 stars 8 forks source link

Some addresses that have had "issues" #11

Open alankjackson opened 5 years ago

alankjackson commented 5 years ago

I'm working through a large list of addresses supplied by the census bureau geocoder, and found a few special cases that the directional prefix/suffix parser has had problems with. 1441 EAST, HOUSTON, TX, 77007 7124 AVENUE E, HOUSTON, TX, 77011 9315 E AVE N, HOUSTON, TX, 77012 1630 DALLAS ST, SOUTH HOUSTON, TX, 77587 (this one I didn't have "South Houston" in the city dictionary, so it got confused at the city stage)

chris-prener commented 5 years ago

This is super helpful, @alankjackson - we can patch the directional issues for sure. With the final one, did everything work once you added SOUTH HOUSTON to your city dictionary? There is going to be trial and error with the dictionaries for sure...

alankjackson commented 5 years ago

Yes indeed, getting that city in fixed it. BTW, I'm testing with 55,000 addresses and performance is fine.

chris-prener commented 5 years ago

Wonderful to hear @alankjackson on the performance front. I've been testing it on a much smaller data set (1,800) and have been pleasantly surprised.

We'll have to do a bunch of patching to close this issue - but I'm on it.