Is it time for Canada to have a national address strategy?
Civic addresses are necessary for many vital applications such as vehicle routing, package delivery, elections, census, emergency response, taxation, public works, GIS and a host of other applications. While several authoritative address databases exist in Canada, they are mostly limited to provincial level data and most often lack format commonality for interoperability. Also, the quality of the content of these databases is sometimes suspect. Is it now time to develop a national address strategy and an address repository for Canada?
Civic addresses are vital to locating property, buildings or people. They also help us identify where one lives or works. A lot of public services, including mail delivery, assessment, taxation, elections, emergency services and driving permits, depend on civic addresses. Many private sector services also depend on accurate civic addresses; for example, hydroelectric services, real estate sales, taxi services, package delivery and banking.
In Canada, addresses are managed by local governments. Municipalities assign names to areas and streets, and numbers to buildings and houses. Efficient address management requires appropriate tools, databases and staff time to keep the address registry error free and up to date. Some local governments, such as Toronto and Vancouver share their addresses through Open Data portals.
The City of Toronto’s One Address Repository dataset provides a point representation for over 500,000 addresses. Each address point is described with a series of attributes including street number, street name, address type, feature class and real-world coordinates.
This image shows the precise position of a few Toronto addresses as red dots. Also shown are the detailed attributes of one of the addresses. This data can be downloaded as a shapefile from the City of Toronto’s open data site, which currently shares 524,679 address points.
This image shows the exact position of a few addresses in Vancouver as red dots. Also shown are the attributes of one of the addresses in the pop-up window. This data can be downloaded as a shapefile from the City of Vancouver’s open data site, that currently shares 100,857 address points.
This seems like a nice and tidy address management process here in Canada. Why then do we need a national address strategy? There are many reasons for an address strategy.
First of all, not all address databases are publicly available. It means that anyone who needs an address database would have to create their own. This results in unnecessary duplicate work. The Cities of Toronto and Vancouver plus a few others have made their address registry available as open data. Why haven’t most other Canadian jurisdictions?
Secondly, not all databases are well developed and properly maintained, which leads to poor data quality. Thirdly, there is no standard address format for collecting and exchanging this data. With Canada’s CRTC recently mandating the use of the National Emergency Number Association’s (NENA) Civic Location Data Exchange Format (CLDXF) for sharing address data for the Next Generation 9-1-1 system, this is an ideal time to look at the CLDXF as a common national interchange format for all Canadian addresses.
Address quality
For most civic addresses in North America, there are address constraint rules that must be followed to maximize data quality. For example, all addresses must be unique; each site must have a primary access point; each unit number must be a positive number greater than one; the parity of the numbers must be consistent with even unit numbers on one side of the street and odd unit numbers on the other side of the street; and unit numbers should increase in one direction along the block face, and so on. However, there are always address exemptions, such as buildings or properties not on a street (for example, properties with only water or air access) that need to be managed as well.
Given that there are somewhere between 14 and 17 million addresses in Canada, it’s not a big surprise that some are in error. Even if 95 percent of the addresses are perfectly correct, it still leaves nearly a million questionable addresses with issues such as missing or duplicate address, incorrect building number, building no longer exists, wrong street, spelling mistakes, accented or special character issues, and a host of other potential inaccuracies.
Moreover, provinces tend to have their own specifications for addresses. Some federal organizations also have their own address database guidelines. For example, Canada Post Corporation follows certain specifications for mailing addresses. The emergency response community also follows certain standards for an address database, which will be NENA CLDXF in the near future. Given that organizations have varying requirements when it comes to addresses, there ends up being plenty of address databases using different standards.
Geocoding
An address by itself is important, but to be really useful, a geographic location for each address is required. The geographic location of the address is called the geocode and it facilitates the search and display of the location of the property or building on a map. When geocodes are included as part of each address, they too could be in error (in other words, the geographic location is incorrect) even if the address itself is syntactically and semantically correct.
For example, here’s the civic address for the building at the famous intersection of Portage and Main, which is at 375 Main Street, Winnipeg, Manitoba. The position of the building is displayed on the map because the address has a geocode field (the latitude, longitude coordinate of the building are included as an attribute of the address).
Mailing or postal addresses such as post office box numbers or rural route numbers are not civic addresses. While postal codes can be useful for helping locate addresses, they are not officially part of the civic address. Postal codes are part of the mailing address and they are considered geographic area identifiers that can be helpful in making sure that the geocode for an address is within the postal code boundary.
The location of a geocode is a complicating factor towards efficient address management. If the address is for the property, the geocode location is often the centroid of the property, but sometimes the entrance to the property is used. If the address is for a building, the geocode location could be the centroid of the building or the front entrance to the building.
The image above depicts the differences in locating buildings and parcels. Shown are: parcel centroid locations (red), building centroid locations (green), building entrance locations (blue), and property entrance locations (purple). Not only is there a substantial distance between these points for large properties that aren’t near the property centroid but even the location points for small urban lots can also show significant distance.
Considering the kind of issues that can happen, address management can be a complicated and resource-intensive undertaking. If several organizations are performing address management for buildings and properties in the same geographic location, that is a tremendous waste of resources. This duplicate time and effort could be used to improve the quality and currentness of the data if the organizations collected, shared and used the same address data.
The Next Generation 9-1-1 has mandated the use of a common address exchange standard in Canada by 2020. Is it time for the remainder of the address management community to move to the new standard and make their data open? Perhaps, the Canadian geospatial community should start working towards developing a national address strategy with the vision to create a common national address registry.