GEOG 5222 Project 4

Geocoding Addresses of Customers Who Performed Home Radon Tests

by Tom Wells -  July 2003

Address Locations

Additional Addresses




ArcView supports "Geocoding Services" that can be used to map a point location based on only a street address.  This project maps the addresses of imaginary home radon tests over a thematic map of "Radon Potential".

Project 4 was described in four parts.

  1. Supplied data setup & Batch Geocoding,
  2. Performing an Interactive Rematch on Unmatched Records,
  3. Analyzing Radon Potential in Soils and Rock Geology, and
  4. Locating a Single Address.

Geocoding, both batch and interactive mode, are introduced in lesson 4.  Step 2 demonstrates problems with bad data and how to manually compensate.  Radon Potential analysis includes an interesting geoprocessing union operation.  Step 4 shows how to locate a single address on a thematic map.

A supplied address table/file with 118 records/addresses was geocoded.  The batch mode geocoding in lesson 4, part 1 was able to locate 114 of the addresses using the supplied geocoding service (roads.mxs).  The 4 unmatched addresses were located interactively in part 2 of the lesson by fixing some typical typos.

The radon potential thematic layer was created by performing a union (mathematical, not labor) on two supplied shape files representing soil and rock.  Both attribute tables contain potential radon ratings on a scale of 1 to 4.  The union geoprocessing combined the 42 soil polygons with the 343 geology/rock polygons to create 823 soil & rock combination polygons (due to the intersections of differing polygonal shapes).  The soil and rock potential radon ratings were simply added together and then normalized to represent low, medium and high radon potential regions. 

Figure 1:    Radon Potential, Roads and Geocoded Street Addresses map.  An individually located address (at 203 Saw Mill Road) was located and turned into a blue star on the map to report that it is in a relatively high potential radon risk area.  The 15370 zip code belongs to Greene county, Pennsylvania but the map apparently comes from elsewhere because this is a fictional data set.  (The geology looks like central Pennsylvania.)  The map represents approximately 40 miles horizontally.


An additional 6 records were downloaded in a single table/address for geocoding.  One address would not match until the zip code was corrected.  The street was valid but in the wrong zip code. 

Figure 2:    Radon Potential, Roads and additional (optional) Geocoded Street Addresses map.  The map represents approximately 40 miles horizontally.


I imagine that rock and soil risk factor addition is a very crude approach to radon potential.  The most famous case was due to a house built on rock: 

"Stanley Watras was living in Boyertown, near to Philadelphia, and was an employee of the Limerick Nuclear Power plant in Pennsylvania. The home of the unfortunate Watras was built on an excavated vein of uranium in the Reading Prong, a Precambrian rock body relatively high in uranium. During December 1984, Watras set off radiation alarms on his way into work. This happened every working day for two weeks and he was regularly decontaminated while the authorities attempted to find the radioactive source. Eventually, the source was found to be radon gas in his domestic dwelling, not a discharge at the plant itself. The levels of radon in his home appeared astonishing to those monitoring at the time, being in the region of 100 000 Bq m-3, even though today levels of greater than 250 000 Bq m-3 are known (Ennemoser et al 1994). The risks associated with living in that house were estimated to be equivalent to smoking 135 packs of cigarettes per day."

Rock and soil (sometimes called unconsolidated rock by geologists) are closely related so I imagine that they can both produce radon gas but the interaction of soil on bedrock is not a mater of simple addition.   For example, a thick deposit of virtually impermeable clay would tend to seal even the most radioactive rock.  


ArcGIS Desktop Help v8.3 (Compiled HTML Help files dated 11/2002 through 1/2003)

GEOG 5222 CD-R, WCGIS\Lesson4\ Dataset 

Jay Computer Services Zip and Area Code Lookup,, accessed 30 July 2003

Optional Dataset (new customer addresses table),,  Downloaded 29 July 2003

The Radon Story at the University of Bradford (England), accessed 30 July 2003.

In this project the geocoding service was provided.  Therefore we did not need to pick the correct style that could work with our data set..   ArcView 8.3 comes with several predefined geocoding service styles designed to work with common address data formats. 
  • Single field geocoding service style
  • US One Address geocoding service style
  • US One Range geocoding service style
  • US Streets geocoding service style
  • ZIP geocoding service style
  • ZIP+4 geocoding service style
  • ZIP+4 Range geocoding service style
  • US Alphanumeric Ranges geocoding service style
  • US Hyphenated Ranges geocoding service style
  • US Cities with State geocoding service style
  • World Cities with Country geocoding service style

According to the ArcMap "Geocoding Services Manager", the provided service (roads.mxs)  is in the "US Streets with Zone (ArcView Style)" geocoding service style.  Right clicking on roads.mdx in ArcCatalog and selecting properties lets the operator see the "Geocoding Service Properties" form.  This dialog form includes a "Fields" section that shows how the reference data (roads.shp) is mapped for this geocoding service.  The following mappings are included.  Data examples or type (address numbers are integer fields) are listed as well.

Geocoding Fields


(data examples or type)
PreDIr FDPRE (W, E, N, S or null)
StreetName FNAME (State Hwy 504, etc.)
StreetType FTYPE (Way, St, Rd, etc.)
SufDir FDSUF (W, E, N or null)
LeftFrom LEFTADD1 (address number)
LeftTo LEFTADD2 (address number)
RightFrom RGTADD1 (address number)
RightTo RGTADD2 (address number)
LeftZone ZIPL (5-digit zipcode)
RightZone ZIPR (5-digit zipcode)

The first four field names are actually mapped three times each to different Geocoding Fields with similar names.  Apparently the service can support two aliases for each of these character fields.

Go To the Top of Page