Geocoding

Aus Geoinformation HSR
Zur Navigation springen Zur Suche springen

The geocoding (geolocalization, DE:Geokodierung) is the process of finding associated geographic coordinates (e.g. latitude/longitude or northing/easting) from other geographic data, such as addresses (street name and street or building number), postal codes or place names. Since address data is a laborious data (like most geospatial data) and often copyrighted there is no single free solution or service.

Geocoding Data

Geocoding Systems

There are many geocoding systems which try to refer to location without coordinates and without streetnames or housenumbers: See https://en.wikipedia.org/wiki/Geocoding .

Open Location Codes (OLC) are a promising way of encoding location into a form that is easier to use than latitude and longitude. They are designed to be used like street addresses, especially in places where buildings aren't numbered or street names do not exist. OLC stands for the technology including the publicly available specs and open source libraries. When speaking about a specific OLC code one can also call it "Plus code" or Short Code.

Example: "bQm Kulturcafé & Bar":

  • Lat/Lon: 47.375938,8.546687
  • Plus Code (complete 10 codes): 8FVC9GGW+9M
  • Plus Code (short code): 9GGW+9M
  • Plus Code (shorter code with geoname): GW+9M Zürich

 >> The section below is outdated and needs to be updated <<

Given a local spreadsheet with not-yet-geocoded address data, there are following current solutions:

  1. ArcGIS Online or Google Fusion Tables (data source for both: Google): no programming, webbased service, data needs to be uploaded, needs Google or AGOL account => See ArcGIS Online and Esri Help Pages.
  2. Google Docs with App Script (data source: Google): predefined scripts, webbased service, data needs to be uploaded => See https://github.com/mapbox/geo-googledocs .
  3. Mapquest (data source: Nominatim/OpenStreetMap): predefined scripts, webbased service, data remains locally => See http://schoolofdata.org/handbook/recipes/geocoding/ .
  4. SpatiaLite (data source: MOPublic/Swiss Cadastre and Statistics): local sw, local data, predefined scripts. See SpatiaLite.

Geocoding im Web

Geocoding APIs

Geocoding for all: A subtle introduction to Geocoding with Taras Dubrava April 2022 https://www.youtube.com/watch?v=lbq7Y2lkVGM

List of geocoders with conditions (price, plans, restrictions): https://getlon.lat/

>> Nominatim API https://nominatim.org/ <<
  • Online batch geocoder for free (for journalists?) geocode.localfocus.nl.
  • OpenCage Geocoder [1]
  • MapQuest Open Geocoding API Service ([2]) (und weitere [3]) (Beschränkungen: Gibt nur Ortschaft/Strasse zurück?; Lizenz: CC-BY-SA)
  • Yahoo BOSS Geo Services (kostenpflichtig!) 'Yahoo! PlaceFinder API' Service (Beschränkungen: mit API-Key; max. 50'000 Requests/Tag; Lizenz: ???)
  • 'Bing Maps APIs' (Beschränkungen: mit API-Key? unübersichtlich; Lizenz: ???).
  • Geoadmin API (Beschränkungen: verschiedene; ?; Lizenz: Bund...???)
  • Geonames findNearestIntersectionOSM oder postalCodeSearch (Beschränkungen: availability; Lizenz: "free"/CC-BY-SA)
  • 'Google Maps Geocoding API' Service (Beschränkungen: nur 2'500 Requests/Tag und Resultat muss im Web sichtbar gemacht werden)
  • tinygeocoder.com:
    • Geocoding "Oberseestrasse+10,Rapperswil-Jona,Switzerland": [4]
    • Reverse geocoding of "47.2230647,8.8165114": [5]
  • GeoAdmin API 2.5, SwissSearch: geocoding and SwissSearch: reversegeocoding [6]

Geocoding mit MS Excel

Geocoding von Adressen aus Excel mit VB programmiert

  • Juice Analytics: Geocodiert einzeln/mehrere Adressen international (Basis Yahoo Webservice) sowie Google Earth ()-Output. Hinweis: Dazu muss man a) einen Yahoo-Account haben und b) eine Application-ID (Authentisierungs-Typ Generic) beziehen; beides ist kostenlos. zum Download...
  • Weiterentwicklung mit Google Maps (falls mehrere Adressen codiert werden sollen, yahoo_id = "YahooDemo" ersetzen durch eigene Application-ID): zum Download...

Geocoding mit QGIS

Siehe auch:

Geocoding mit Python

Siehe auch: Python.

"Géocoder en masse plusieurs milliers d’adresses avec Python et Nominatim" par AdrienVH, le 18/01/2015 [7].

 """ ====================================================
 geocode.py - a geocoder based on OSM / Nominatim
 Dependencies: GeoPy
 Input       : File adresses.csv
 Output      : File resultat.csv
 ==================================================== """
 import csv
 from geopy.geocoders import Nominatim # needs GeoPy lib!
 geocoder = Nominatim()
 inputFile = open('adresses.csv', 'rb')
 outputFile = open('resultat.csv', 'w')
 try:
   outputData = csv.writer(outputFile, delimiter=';', lineterminator='\n')
   outputData.writerow(('adresse', 'latitude', 'longitude'))
   inputData = csv.reader(inputFile, delimiter=';')
   for ligne in dataInput:
     adresse = ligne[0] + ", " + ligne[1]
     try:
       location = geocoder.geocode(adresse, True, 30)
       outputData.writerow((adresse, location.latitude, location.longitude))
     except Exception as inst:
       print inst
 finally:
   inputFile.close()
   outputFile.close()

Geocoding with PostgreSQL and Python

 CREATE LANGUAGE plpythonu;
 CREATE OR REPLACE FUNCTION geocode(address text)
 RETURNS text AS $$
 """Returns WKT geometry which is EMPTY if address was not found. 
 Please respect the Nominatim free usage policy: 
 https://operations.osmfoundation.org/policies/nominatim/ 
 """
 import requests
 from time import sleep
 base_url = 'https://nominatim.openstreetmap.org/search?'
 params = {"q": address, "format": 'json', "limit": 1, 'User-Agent': 'plpythonu geocode hack'}
 response = requests.get(base_url, params=params)
 loc = response.json()
 if len(loc) == 0: 
     wkt = 'SRID=4326;POINT EMPTY'
 else:
     wkt = 'SRID=4326;POINT(%s %s)' % ((loc[0]['lon'], loc[0]['lat']))
 sleep(0.1)
 return wkt
 $$ LANGUAGE plpythonu VOLATILE COST 1000;
 -- Test:
 SELECT ST_AsGeoJSON(geocode('Vulkanstrasse 106, Zürich'));
 SELECT ST_AsGeoJSON(geocode('Weihnachtsmann'));
 SELECT ST_GeomFromEWKT(geocode('Vulkanstrasse 106, Zürich'));
 -- Geocode!
 WITH my_customers_tmp(id, address) AS (
   (values 
    (1,'Vulkanstrasse 106, Zürich'),
    (2,'Bellevue, 8001 Zürich'),
    (3,'Weihnachtsmann')
   )
 )
 SELECT id, address, geom, NOT ST_IsEmpty(geom) as found 
 FROM (
   SELECT id, address, ST_MakeValid(ST_GeomFromEWKT(geocode(address))) AS geom 
   FROM my_customers_tmp
 ) as tmp;

Weblinks

Generell/Liste: http://del.icio.us/search/?all=geocoding