The aim of OpenAddresses.org is to establish a copyright-free database of geocoded addresses, using donations and voluntary help. Such a database has a wide range of applications including spatial analysis and geomarketing. The database is intended to be of greater accuracy than navigation systems which typically have insufficient accuracy for microgeographic analysis.
OpenAddresses includes a user web interface and a number of REST services. Data is typically hand-digitized or donated by institutions and public authorities.
The OpenAddresses project was started in 2007 by the Institut Vermessung und Geoinformation at University of Applied Sciences Northwestern Switzerland (FHNW). In early 2010, the project was launched as a new application at OpenAddresses.org. The team has grown and includes members of FHNW, camp2camp, Fh Karnten, ubiThere, and Inser. The user interface has been redesigned using MapFish, and incorporates a number of REST services for geocoding and reverse geocoding.
OpenAddresses publishes data under the CreativeCommons license. Philosophically has much in common with the OpenStreetMap (OSM) project, and OpenAddresses periodically synchonizes its data with OSM.
Geolocated addresses are used in a wide range of geospatial and geoweb applications. Existing datasets typically have restrictive rights, high cost, inadequate accuracy, or lack of coverage. OpenAddresses attempts to fill this gap. If the OSM example is anything to go by, it will be particularly valuable in areas of low potential profitability – eg. rural areas and the Third World.
Individuals can help the OpenAddresses.org project by digitizing point addresses using the Mapfish-based interface at OpenAddresses.org:
Note the pop up window including a helpful tool tips. To use the interface, simply click on the map at the new address location, and enter the address information in. There are still a few teething problems – I received a “data write error”. This could have been a legitimate problem, but it would have been useful if the error message was as informative as the tool tips.
The database benefits even if people only enter the the address locations for themselves and people they know. This is coordinate / street address – so you and your friends cannot be identified. Donations of aerial photographs, building plans, and WMS tiles also help other users to locate addresses.
As well as the user interface, the MapFish framework is also used to implement the OpenAddresses.org REST services. Data fields include street address information, a quality field (currently ‘Digitized’, ‘GPS’, or ‘Linear interpolation’), and coordinate information. The REST interface can work with geospatial ‘box’ queries and text queries. It also supports different coordinate systems using EPSG codes. Complex queries can be constructed, for example “Return the first three addresses in city X and return only the street, housenumber, and city attribute without any geometry“; or “get upto 5 addresses in Geneva that contain the word ‘Vogt’“. Reverse geocoding queries work in a similar manner and can query all addresses within a certain radius of a specific coordinate.
OpenAddresses.org shows a lot of promise to be an important project for both desktop and geoweb applications. The biggest problem is the same as for most of these community projects – including OpenStreetMaps, and that is of critical mass. To really work (and to attract volunteers), it really needs a significant amount of data and coverage. OpenStreetMap was able to seed its database with free data sources such as TIGER/Line. OpenStreetAddresses.org has managed to attract a number of government donations including the recent donation of address data by North Carolina. It needs to attract more such data donations. Perhaps the British Royal Mail could donate its PAF database, following in the footsteps of the Ordnance Survey’s recent data releases?
Another problem could be that of address format. US address formats are very formal. Many European addresses are less formal, and many other countries could best be described as “informal”, eg. “200m north of the SuperKike in Town Y” would be a valid (and official) costa Rican address! OpenStreetAddresses could be very attractive in the Third World, where this problem will be more acute than the vast majority of geocoding applications that we have discussed here at GeoWebGuru.
This article is primarily edited from a number of articles published by the OpenAddresses.org organization and used with permission.