April 2020
« Mar    

Geocoding – Address Cleanup Methods

Geocoding, or converting address information into XY coordinates, has a number of subleties to it, so if you have reached this page you probably have some experience with this, and may be struggling with failed geocoding runs.

This page is focused on attempting to deal with BAD addresses.  The County’s CAMS system generally handles correct addresses well, but many large files deal with addresses that have been incorrectly entered, with bad spacing, misspelled street names, strange unit numbers, etc.

Richard Ledesma from our Public Social Services Agency (DPSS) has been working with this situation, and has written a number of scripts that can help solve certain types of issues.  He was given a table of 546,000 addresses, of which 50,000 were not geocoding.  He developed some pattern recognition as well as ColdFusion applications to help fix these addresses in an automated fashion

Click here to read his background paper on this.

Along with Richard, I will use this page to try to provide tools and instructions on how to deal with bad addresses.

Method 1: Using the LA County Standardization Rules

standardizingIn this post (LA County Locators and Matching Rules) I show how to install and use the County locator rules.  A hidden tool for fixing addresses is the “Standardize Addresses” function in ArcMap.  This will take every address in a source file, then attempt to break it into its constituent parts (house number, pre direction, street name etc).  The standardization can help you find and fix common errors, and then you can concatenate the correct fields back together and try to regeocode.  It’s also a powerful way to organize your addresses, and begin to inspect them in programs like Excel where you can establish where your biggest error sources come from (or where our locators fail).


Method 2: Using the DPSS Address Cleanup Scripts

There are other methods for cleaning up addresses.

Currently you need to have a ColdFusion server to use them, but Richard will be working to port them to either Python, PHP, or both, so we can provide this as a service.

You can download these scripts below.

To download the clean up scripts in text format, click here


If you have questions or comments, please send us an email and we will do our best to respond.