Convert lat/lon to ZIP/Postal code using Python

Question

I have this dataframe:

      Lat       Lon
 29.39291 -98.50925
 29.39923 -98.51256
 29.40147 -98.51123
 29.38752 -98.52372
 29.39291 -98.50925
 29.39537 -98.50402
 29.39343 -98.49707
 29.39291 -98.50925
 29.39556 -98.53148

I want to convert these lat/lon rows to ZIP codes (for each row), using Python.

Answer 1

You can use geopy and its Nominatim geocoder.

Here is an example using the DataFrame you provided:

import geopy
import pandas as pd


def get_zipcode(df, geolocator, lat_field, lon_field):
    location = geolocator.reverse((df[lat_field], df[lon_field]))
    return location.raw['address']['postcode']


geolocator = geopy.Nominatim(user_agent='my-application')

df = pd.DataFrame({
    'Lat': [29.39291, 29.39923, 29.40147, 29.38752, 29.39291, 29.39537, 29.39343, 29.39291, 29.39556],
    'Lon': [-98.50925, -98.51256, -98.51123, -98.52372, -98.50925, -98.50402, -98.49707, -98.50925, -98.53148]
})
zipcodes = df.apply(get_zipcode, axis=1, geolocator=geolocator, lat_field='Lat', lon_field='Lon')

>>> zipcodes
0    78204
1    78204
2    78204
3    78225
4    78204
5    78204
6    78204
7    78204
8    78225
dtype: object

Answer 2

For geocoding with ArcGIS, credentials have to be provided gis = GIS("http://www.arcgis.com", "username", "password").

---

In terms of pandas, this code should do the work (adapted from @Marcelo Villa's answer)

from arcgis.geocoding import reverse_geocode
from arcgis.geometry import Geometry
from arcgis.gis import GIS
import pandas as pd

gis = GIS("http://www.arcgis.com", "***", "***")

def get_zip(df, lon_field, lat_field):
    location = reverse_geocode((Geometry({"x":float(df[lon_field]), "y":float(df[lat_field]), "spatialReference":{"wkid": 4326}})))
    return location['address']['Postal']

df = pd.DataFrame({
    'Lat': [29.39291, 29.39923, 29.40147, 29.38752, 29.39291, 29.39537, 29.39343, 29.39291, 29.39556],
    'Lon': [-98.50925, -98.51256, -98.51123, -98.52372, -98.50925, -98.50402, -98.49707, -98.50925, -98.53148]
})

zipcodes = df.apply(get_zip, axis=1, lat_field='Lat', lon_field='Lon')

---

Probably not as smart as pandas but using Python's internal module csv and ArcGIS's API for Python Reverse Geocoding.

from arcgis.geocoding import reverse_geocode
from arcgis.geometry import Geometry
from arcgis.gis import GIS
import csv

gis = GIS("http://www.arcgis.com", "***", "***")

coords = [
(29.39291, -98.50925),
(29.39923, -98.51256),
(29.40147, -98.51123),
(29.38752, -98.52372),
(29.39291, -98.50925),
(29.39537, -98.50402),
(29.39343, -98.49707),
(29.39291, -98.50925),
(29.39556, -98.53148)
]

result = []

for lat, lon in coords:
    pt = Geometry({
        "x": float(lon),
        "y": float(lat),
        "spatialReference": {
            "wkid": 4326
        }
    })
    try:
        result.append({'lat': lat, 'lon': lon, 'geocoded': reverse_geocode(pt)})
    except:
        pass

result_zip = []

for item in result:
    result_item = {
        'lat': item['lat'],
        'lon': item['lon'],
        'zip': item['geocoded']['address']['Postal']
    }
    result_zip.append(result_item)

keys = result_zip[0].keys()

with open('output.csv', 'w', encoding='utf8', newline='') as output_file:
    dict_writer = csv.DictWriter(output_file, keys, delimiter=';', quoting=csv.QUOTE_NONE, lineterminator='\r')
    dict_writer.writeheader()
    dict_writer.writerows(result_zip)

Output csv-file looks as following