Objective: Using R, get lat/long data for a vector of addresses through Google Maps. Approach: I began with the excellent code from Tony Breyal at: Geocoding in R with Google Maps and modified it slightly to suit my own purposes (I need the function to return a SpatialPoints object with all of the addresses in it and the projection string defined).
So we have:
library(RCurl)
library(RJSONIO)
library(maptools)
#unchanged from Breyal's code
construct.geocode.url <- function(address, return.call = "json", sensor = "false") {
root <- "http://maps.google.com/maps/api/geocode/"
u <- paste(root, return.call, "?address=", address, "&sensor=", sensor, sep = "")
return(URLencode(u))
}
#obviously this would be better vectorized, but I left it in this form to help me
#understand where the errors were coming from
batchGeoCode <- function(address.list) {
lat<-vector(mode="numeric",length=length(address.list))
lng<-vector(mode="numeric",length=length(address.list))
for(i in 1:length(address.list)){
address<-address.list[i]
u <- construct.geocode.url(address)
doc <- getURL(u)
x <- fromJSON(doc,simplify = FALSE)
#I put the try in here just so I could see if it was always getting
#hung on the same address or the same special character or the like
errtest<-try(x$results[[1]]$geometry$location$lat)
if(is(errtest, 'try-error')){
lat[i]=0
lng[i]=0
}else{
lat[i] <- x$results[[1]]$geometry$location$lat
lng[i] <- x$results[[1]]$geometry$location$lng
}
}
points.df<-data.frame(lat,lng)
coordinates(points.df)=~lng+lat
points.sp<-SpatialPoints(coords=points.df,
proj4string=CRS("+proj=longlat +datum=WGS84"))
return(points.sp)
}
#some sample data that is enough to cause a fail for me--groceries and mini marts in
#South Seattle. All of these return a valid address when entered individually in
#Google Maps
address.list<-c("T W YOUNG MARKET 7144 BEACON AVE S SEATTLE WA 98108",
"RAINIER MARKET CTR 3625 1ST AVE S SEATTLE WA 98134",
"JING JING ASIAN MARKET 12402 SE 38TH ST BELLEVUE WA 98006",
"HILAL MARKET PLACE 15061 MILITARY RD S SEATAC WA 98188",
"GUADALUPE MARKET 1111 SW 128TH ST BURIEN WA 98146",
"JAMAL MARKET 14645 TUKWILA INTERNATIONAL BL TUKWILA WA 98168",
"C & D INTL MARKET SEATTLE WA 98198",
"GUERRERO MARKET 17730 AMBAUM BLVD S # B SEATTLE WA 98148",
"MI CASA MARKET 17174 116TH AVE SE RENTON WA 98058",
"GUERRERO MARKET 17730 AMBAUM BLVD S # B SEATTLE WA 98148",
"MI CASA MARKET 17174 116TH AVE SE RENTON WA 98058",
"BILISEE MARKET 8300 RAINIER AVE S SEATTLE WA 98118",
"ALI SEATAC INTL MARKET 16324 INTERNATIONAL BLVD SEATAC WA 98188",
"OCEAN MARKET 2119 RAINIER AVE S SEATTLE WA 98144",
"RED APPLE MARKETS 9627 DES MOINES MEMORIAL DR SEATTLE WA 98108",
"ANKGOR MARKET 9660 16TH AVE SW SEATTLE WA 98106",
"SIXTEENTH AVENUE GROCERY 9001 16TH AVE SW SEATTLE WA 98106",
"RAINIER ORIENTAL MARKET 9237 RAINIER AVE S SEATTLE WA 98118",
"CLEAN MACHINE LAUNDRYMAT SEATTLE WA 98118",
"HALIN GROCERY & DELI SEATTLE WA 98118")
#call the function
result<-batchGeoCode(address.list)
With short lists this works, but as the list gets over 15 I start getting 'no result' responses. I am pretty sure this is a throttling problem (such as Slow down Geocoding a batch of Address), but not too clear on how to implement this within R so it fits the specific needs of this function.
Thanks in advance,
Chris
