Using Geopy for Spatial Data Analysis

Geopy is a useful python package to deal with spatial data, such as locating the coordinates of addresses, cities, countries, and landmarks or reverse. In the Two Sigma Connect Competition in Kaggle, I found that some of the given latitude longitude do not match the addresses, so I use Geopy to clean the mismatched latitude and longitude of rental listings.

Import Packages and Read Data

I am working with rental listings in New York City, the addresses should be in NYC, so I append 'New York City' to the addresses ensure its accuracy.

from geopy.geocoders import GoogleV3
bad_location['street_address'] = bad_location['street_address'] + ',New York City'

Geoencoding

There are several geolocation services we can use, such as Google Maps, Bing Maps, or Yahoo Boss. I choose to use Google Maps Geoencoding API (GoogleV3) because it is insensitive to missing information. Using a standard GoogleV3 API, we can have 2,500 free requests per day and 50 requests per second. Before using it, we need to register to get a free API key following this link.

def do_geocode(address):
    geolocator = GoogleV3(api_key = 'your_api_key')
    try:
        location = address.apply(geolocator.geocode)
        latitude = location.apply(lambda x: x.latitude)
        longitude = location.apply(lambda x: x.longitude)
        return (latitude, longitude)
    except GeocoderTimedOut:
        return do_geocode()

(latitude, longitude) = do_geocode(bad_location['street_address'])
bad_location.loc[:,'latitude'] = latitude
bad_location.loc[:,'longitude'] = longitude

Quite simple, within a few minutes, we can generate the geolocation for the addresses.