Code to do the geocoding of the foodborne addresses.

Geocode foodborne addresses

library(RCurl)
library(RJSONIO)
library(ggmap)

df <- read.csv("submissions-2013-11-05.csv", stringsAsFactors=F)
names(df) <- tolower(names(df))
# order by id, or basically date
df <- df[order(df$id),]

df$lat <- df$lng <- rep(NA, nrow(df))
# do this in a loop since we have to sleep to avoid google's rate limiting
for(i in 1:nrow(df)){
  rest.address <- ifelse(grepl("chicago", df$restaurant.address[i], 
                               ignore.case=T), df$restaurant.address[i],
                         paste0(df$restaurant.address[i], ", Chicago IL"))  
  results <- geocode(rest.address)
  df$lng[i] <- results$lon
  df$lat[i] <- results$lat
  Sys.sleep(.35) # can't ask google for these too fast...
}