Geocoding: Fundamentals, techniques, commercial and open services
Last modified: 2010-07-23
Abstract
Referencing information and features to geographic locations is an essential component of daily live. In newspapers and television news we are accustomed to see maps illustrating the location of events in politics, sports, or natural hazards. The process of assigning location to such features is called georeferencing.
Postal addresses, a specific way of locational description, are the essential means by which people express their location in the real world. Also many features in modern societies in administration and business are related to addresses which are used as unique identifiers to geographical locations, often expressed by a pair of longitude and latitude values. Although addresses are data containing location, they do not contain coordinates. Therefore an important feature in many applications especially in geographic information system (GIS) is the capability to locate addresses, i. e. to geocode to address level.
During the last few years there has been also a significant progress to geocode addresses by commercial Internet mapping application programming interfaces (APIs) and Internet services. Recent advances, mainly in the field of Internet technology, have encouraged people to develop and to integrate addressing data models from variety of addressing system. It has made geocoding to be more popular, and using such online tools, the geocoded address can be retrieved and displayed. The process of geocoding is not perceived as an individual service by the user but as a standard function of WebMapping.
In the beginning of this paper, fundamentals of address schemas and of geocoding approaches are introduced. The process of parsing addresses and matching with reference datasets is explained. Phonetic algorithms including Soundex and Kölner Phonetics, and Levenshtein distance are explained in this context.
Some open solutions available on the web are reviewed and compared concerning their coverage, the APIs offered, and the response formats:
Full Text: Paper (PDF)