Python script for geocoding, adapted from http://rowanwinsemius.id.au/blog/geocoding-addresses-from-a-csv-with-python/

geocodescript.py

#import the csv module which we'll need to read and write our csv's
import csv

#import the shutil module which we use to create a copy of our original csv
import shutil

#import the Nominatim module from geopy, a bit more on this in the text below
from geopy.geocoders import Nominatim
geolocator = Nominatim()

#Let's start by creating a copy of our raw file just to be safe
shutil.copyfile("dirtyrestaurants2.csv","dirtyrestaurants2_processed.csv")

#open the file and assign it to the var f
f = open('dirtyrestaurants2.csv')

#create a blank array called addresses
addresses = []

#use the csv reader to actually start doing stuff
csv_f = csv.reader(f)
# iterate through each row, and only grab column 1, and add it to the address array
for row in csv_f:
   addresses.append(row[1])
   # print addresses

#close our original file
f.close()

#create a new empty array called locations
locations = []

#for each value in the addresses array that we created earlier geocode it and save it to a variable called location, we'll then append that single location to the locations variable
for index,val in enumerate(addresses):
   location = geolocator.geocode([val], timeout=15)
   # print index + 1, location.latitude
   locations.append([location.latitude, location.longitude]) 

   
#Create a header row for our new csv
header = ['Name','Address','Latitude','Longitude']

#open the original csv and grab all the data, place it in a var called data, and close the file again
f = open('dirtyrestaurants2.csv')
data = [item for item in csv.reader(f)]
f.close()

# create a blank arraycalled new_data
new_data = []

# for each item in data append a location, then add the complete item to the new data variable
for i, item in enumerate(data):
    item.append(locations[i][0]) #append latitude
    item.append(locations[i][1]) #append longitude
    new_data.append(item)

#open the new csv and write the header row followed by a row for each object in the new_data array
f = open('dirtyrestaurants2_processed.csv', 'w')
csv.writer(f, lineterminator='\n').writerow(header)
csv.writer(f, lineterminator='\n').writerows(new_data)
f.close()