Monday, December 5, 2016

6 Tips for Preparing Your Data for Geocoding

Geocoding is the process of assigning a latitude and a longitude to a specific address.  It is what allows people to place points on maps.  When you enter that address into Google Maps, it is geocoded and marked on the map for you.  At HealthLandscape we have a tool that allows you to quickly add between 1 and 300 points to your map within most of our mapping tools for free.  Our QuickGeocodes tool can be found by clicking the “Tools” button above the map- just check the checkbox to open the tool.  It is important to note that the QuickGeocodes tool is not HIPAA compliant so you should not be using patient street addresses with this tool.

Let’s say you're ready to import your data set into QuickGeocodes. How can you be sure that your data set is accurate and will geocode correctly? Here are six quick tips:

#1: Be sure you have a complete street address


Problem: Geocoding relies on complete street addresses, including house number, street, city, state, and ZIP Code. The address won't geocode if any part of the address is missing.    

Solution: Verify that every street address is complete.

#2: Make sure each part of the address is contained in a separate field (cell)


Problem: The entire address is contained in one field, as in:


1234 Main Street Cincinnati Ohio 45202  



Or parts are in one field as in:

1234 Main Street
Cincinnati, OH 45202
                  
                  

Solution: Break each part of the address into separate fields:

1234 Main Street
Cincinnati
OH
45202

#3: Notice addresses that don't make sense


Problem: Some addresses have a street name only with no number. Or a number with no street name.

Solution: Scan to see if the addresses make sense. You obviously can't verify every address, but look for those that seem unreasonable. Perhaps someone inserted a phone number instead of a house number.

#4: Eliminate address nicknames


Problem: Those who recorded the data used abbreviations for city or street names. For example, a popular abbreviation for "Los Angeles" is "LA." It's unlikely that any geocoding system will match "LA" to "Los Angeles."

Solution: Go through the data and standardize city and street names.

#5: Know what will geocode and what won't


Problem: Sometimes people substitute a post office box for a true mailing address, or they list both a post office box and a mailing address in the same field. Similarly, they may list a building name in place of a true address. For example, in Cincinnati, they might list an address as "Carew Tower" instead of 441 Vine Street, Cincinnati, Ohio, 45202.

Solution: Eliminate any address that is not a true street address, such as a post office box or rural route address. It won't geocode anyway. And replace building names with actual addresses.

#6: Establish data entry best practices before you collect the data


Keep this tip in mind for your next data collection effort. The cleaner your data, the better your results. Our "Introduction to HealthLandscape" webinar teaches you how to geocode your own data, as well as access data from our Community HealthView data library. Register today using the link below.