Geocoding

PROBLEM: Scenario:

A local commercial establishment in Raleigh NC plans to expand its customer base. So they would like to develop an effective campaign to reach the potential customers in unrepresented areas. To ensure effective planning for their advertisement campaign, the commercial establishment has contacted you, as a GIS professional, to provide an analysis on where their current customers come from and to make recommendations for effective advertisement campaign in areas less-well represented in the Wake County area. 

All assignment data are provided in a zip file (geocode_data.zip) where Wake zip code, Wake streets and Wake boundary data including customers data are found. Wake county shape files are found from US census originally and commercial establishment has provided the customer data with the addresses in excel form.

Analysis Procedures:

First of all by reviewing the customer data, full addresses of 255 customers are found. Wake Zip code vector data has 50 zipnum (only zipcode) with city names including Zip codes are reviewed. There are two geocoding process using two different reference data. By using one reference data like Wake Zip codes all customer addresses has been plotted on the map.

In first process one new address locator has been created from the main Geocode folder in Arc Catalogue. Wake Zip code data is used for reference data. In the address locator style, US address-5 digit Zip has been chosen. After activating new address locator and added the new shape file in Geocoding address locator manager. Then Batch geocoding process executed with customer Excel data with new shape file created in address locator.

After running Geocoding process customer data has been plotted on the map. 35 points on the map represented 255 customers, because of Zip code centroid the points of all customers’ addresses were located in the center. By reviewing and re-matching the errors in the raw data with the help of USPS site, enabled more matching results.

In Second process, one new address locator has been created from main Geocode folder in Arc Catalogue. Wake Street layer is chosen as reference layer and Dual range style has been chosen. For this address locator Geocoding option-“map with no zone – yes” has been chosen. After activating new address locator new locator added to the map and interactively located address on the map by using find tool. In this way four addresses with labelled point has been plotted on the map. Then after running Batch Geocoding process for the Customer address table, it didn’t work out, because street number and street should be together in one column. So I needed to modify customer table. I created a dBase table from Customer table and added to the table of contents. Then by adding a field in dBase table and by using field calculator, Street number and street name added together in this new field. After running batch geocoding process again matched, tied and unmatched addresses found. Then I corrected two errors in the address data through reviewing and rematching, got  final result and exported a map out of it.

PROCESS DIAGRAM:

                                      Workflow Diagram

Results:

Application and Reflection:

Example 1:

The customers from same Zip code recorded in one point. But there are many polygons with same zip code. So, first method is not realistic to find out accurate customers’ spatial location. If there is a disease outbreak in different area or zip code, then I can use zipcode data as a reference layer to find out where the disease outbreak is more prevalent than other areas or zip codes.

Example 2:

On the contrary, second method was reliable to find out actual location. Interactively labelled points are perfectly positioned on the map.

We can help with disaster management by using street data and affected people data by making a map of affected locality. Recently Hurricane Harvey damaged different neighborhoods of Texas. If I have the affected people’s data in XLs format and Street data from Tiger line shape file, I can geocode the affected people’s  data on the map and make a map of affected area. 

Back to top