Geocoding using Google Map API via R with examples

Although I've taken an ArcGIS course before, I still like to use R as much as possible to complete some daily tasks (the reason is obvious...). These days one issue about geocoding or location based analysis comes out, driving more attention from me. Basically the task is to geocode and to check the accuracy of business addresses.

People around me mostly are using ArcGIS or use other expensive tools like TeleAtlas to do the geocoding. As I am not the core GIS person, I would like just to use my usual way --R -- to do the task. Here is the code I learned from this blog "Calling Google Maps API from R". I feel the code that the blog offers works pretty well. So I borrow it for my future reference and made some modification and examples based on my work experience.

For more information about Google Map API, we can refer to google's official documents: https://developers.google.com/maps/documentation/geocoding/

#####################################################################

#  R code to call Google Map API 

#  Source codes: http://svnwang.blogspot.com/
#  Reference: http://statisfaction.wordpress.com/2011/10/05/calling-google-maps-api-from-r/

#####################################################################

library(XML) # use install.packages("XML") if you haven't install this XML library before

getDocNodeVal=function(doc, path)
{
   sapply(getNodeSet(doc, path), function(el) xmlValue(el))
}

gGeoCode=function(str)
{
  library(XML)
  u=paste('http://maps.google.com/maps/api/geocode/xml?sensor=false&address=',str)
  doc = xmlTreeParse(u, useInternal=TRUE)

  lat=getDocNodeVal(doc, "/GeocodeResponse/result/geometry/location/lat")
  lng=getDocNodeVal(doc, "/GeocodeResponse/result/geometry/location/lng")
  formatted_address =getDocNodeVal(doc, "/GeocodeResponse/result/formatted_address")
  location_type = getDocNodeVal(doc, "/GeocodeResponse/result/geometry/location_type")
  status = getDocNodeVal(doc, "/GeocodeResponse/status")
  
  list(lat = lat,  # latitude 
       lng = lng,  # longitude
       formatted_address=formatted_address,  # full addresses suggested by google
       location_type = location_type, # geocoding accuracy
       status=status # status of geocoding: OK or zero_results
       )
}


#Example 1: you have a right address to geocode
str1 = "11 Wall St, New York, NY"
gGeoCode(str1)

#Example 2: your address is too general, 
#           but still google can map it by approximation.
str2 = "Wall St, New York, NY"
gGeoCode(str2)

#Comments: 
#as you can see from the result,there are multiple matched pairs of geocode and matched address.
#in practice, you need to choose which one is more to your need.

#Example 3: your address is not accurate, but google can guess what it is.
str3= "11 Wall Rd, New York" #actually, it should be "Wall St"
gGeoCode(str3)

#Example 4: your address is too bad, far from accurate, thus cannot be geocoded
str4= "1021 Watl P1lz"
gGeoCode(str4)