Canadian Census: Geography Products and Services
the Evolution of their Development and Delivery
Abstract
The development of a national digital geographic base for the 2001 Census has brought fundamental changes to a number of aspects of Census collection and dissemination. This includes the introduction of new and enhanced products and services that are built upon the content and characteristics of the new geographic base. In addition, since the last Census the delivery of geographic information via the Internet has evolved from being an option to the point where it is now the method of choice. The paper reviews the developments made possible by the new base, including product and service creation for the Internet, and future directions that the new base offers.
Introduction
Statistics Canada’s Geography Division develops and maintains a geographic infrastructure to support survey-taking and to facilitate the provision of survey results to the public. Another important responsibility, reflecting the broader Agency mandate, is to make the results of these activities available to the public. In developing geography products and services that support the use of census data, the goal of the Division is to enhance the utility of those statistics at all levels of geographic analysis. With each census, the objective is to improve the geography products in order to serve data users better.
The geographic infrastructure of the Division provides the cornerstone for the collection of the census, our largest survey, and the dissemination of the census results. This infrastructure is used to delineate geographic areas of collection and dissemination, to produce high quality collection and dissemination maps, and to create a variety of other tools related to the census taking process. Geographically referencing dwellings to blocks, small geographic areas that are the responsibility of an individual census representative, provides control in the collection process. At the most basic level, it ensures that the residents of all dwellings in Canada are counted and that no households are counted twice. These same small areas serve as the geographic basis for the processing and release of the results of the census.
The geographic infrastructure, correctly called the spatial data infrastructure, consists of a digital geographic base of census geographic unit boundaries, roads, hydrography and related metadata. This infrastructure has undergone fundamental changes since the 1996 Census. Roads are a key element because they are used in the delineation of the geographic units of the Census and also provide census staff and data users with visible references on maps. Since 1996, the coverage of roads in digital format has been expanded from major urban areas only, to the point where all of Canada is now covered. As a result, for the 2001 Census, this newly expanded content has been a stimulus to bring about many changes in our processes of census collection and dissemination. The design and development of the new spatial data infrastructure, and its impact on the process of collecting the census has been addressed in a number of other academic papers. Given the release of the first 2001 Census results, as well as most of the Geography tools in March 2002, it is now appropriate to examine the extent to which the dissemination of the census has changed as a result of the new base. The objective of this paper, then, is to review the new dissemination developments for the 2001 Census and to examine these developments in the context of past practices.
Past Dissemination of Census Geography
Reference maps allow data users to view the extent of their areas of interest in relation to the geographic units by which census data are made publicly available. These have always been a major output of the geography dissemination program. From early days, these maps have been produced on paper. Until 1996, most of the map series were manually produced. When the Geography Division began to adopt GIS technology in the late 1960’s in advance of the 1971 Census, the creation of maps and source materials using automated techniques began. At the same time, products that put this information in the hands of data users in digital format started to become available. The Area Master File, containing roads and streets for parts of 14 major urban areas, was released in 1969. Also in place following the 1971 Census was a GIS system that facilitated the retrieval of data for user defined geographic areas.
By the 1981 Census, large, automated systems were tested for the preparation of maps of urban areas with 8,000 of these maps at the enumeration area being produced for 1986 Census collection when the system was put into full production. A major outcome from the 1986 production system was the enhancement of road network information to improve its usefulness for automated mapping. The enhanced road network also made possible the creation of an enumeration area boundary file for the first time. Created directly from the road features where coverage was available, this file of 46,000 enumeration areas was disseminated for the first time following the 1991 Census. The enumeration areas served as the building blocks from which all other units in the hierarchy were delineated. This digital boundary file of enumeration areas was then used to create a series of positionally consistent boundary files for census geographic units at all levels of the hierarchy up to province. This gave GIS users a set of integrated boundaries with which to analyse and map census data. It also meant that enumeration area reference maps could be prepared specifically for dissemination. Previously, enumeration area data users wishing to identify the exact boundaries of particular units were forced to consult updated and cleaned copies of the field maps used in the collection process. The enumeration area maps produced for large urban centres in 1991 were created in 11"x 17" format and followed a grid. Index maps were used to locate specific sheets. Clients, however, found the grid approach difficult to use and the series for 1996 were designed to display enumeration areas by census tract. The creation of this 1996 series of reference maps was also marked by the development of an automated system to produce them. Many of the design specifications for this particular series have been retained for the 2001 product although the production system itself has been redesigned and enhanced.
The Spatial Data Infrastructure and Census Dissemination
The development of the new spatial data infrastructure has been achieved in partnership with Elections Canada. Elections Canada requires the ability to produce high quality maps at relatively short notice to support Federal elections. The maps are also used in the redelineation of Federal Electoral Districts. Statistics Canada and Elections Canada recognised efficiencies in combining resources and efforts to develop a common geographic base and have proceeded to do so since 1996.
A major element of the new infrastructure development has been the expansion of the existing road network to cover the country, and the updating of the road network. Information to do this came from the National Topographic Database of Natural Resources Canada, and a variety of other sources including fieldwork. In addition, all existing information, including census boundaries and roads, was moved to North American Datum 83. As was the case in the past, the new base also carries road names.
Benefits of the new base for dissemination are realised in many ways. In the geographic unit delineation phase of our work, the new national digital coverage of the spatial data infrastructure has made it possible to delineate blocks. Blocks are new geographic units for 2001 and replace enumeration areas as the "atomic unit" from which all other units in the census geographic hierarchy are created. Blocks are polygons formed by the intersection of roads. Enumeration areas, the basic unit for the management of census collection and until now the smallest national unit in the hierarchy, are now groupings of blocks. In the past, enumeration areas were also used to disseminate the results of the census. Because they were defined to serve census collection, they had a number of shortcomings when used for data dissemination that led to calls from data users for a better basic geographic unit for dissemination.
With the introduction of the block, it was also now possible to construct a new unit specifically for the small area presentation of data, the dissemination area. Like the enumeration area, the dissemination area is an aggregation of blocks, but in the case of the dissemination area, the criteria used in the aggregation process produce units optimised for geographic data analysis. Enumeration areas in the past were always a compromise between collection and dissemination needs with the balance usually tipped in the direction of collection given the costs associated with that activity. Compared to the enumeration areas, the dissemination areas are more uniform in size and generally cluster around an optimum of 550 persons per unit. Fewer units with very small populations will reduce the amount of data suppressed for reasons of confidentiality.
The new base also offers direct benefits for the products themselves. The availability of a national digital base now means that reference maps for all the census geographic units for all parts of the country can be produced using automated systems. This is the first Census for which it has been possible to use geographic information systems technology to produce all reference maps. Digital files of boundaries for census geographic units have been available for a number of years, however, a national road reference layer, vertically integrated with the boundaries is now available for the first time.
Another major advantage of the national digital base is that products designed specifically for dissemination via the Internet can be developed. Offering Statistics Canada data users greater flexibility in how they can obtain geographic information has been a major element of the 2001 census geography dissemination planning and the option of Internet delivery has been a major component of this.
Dissemination for the 2001 Census
Reference Maps
The number of reference maps series has been expanded for the 2001 Census, all maps have been produced digitally for the first time, and users are now offered a choice when they want to obtain maps. Maps are now available on the Internet for downloading from the Statistics Canada website. Users who wish to have a paper copy of a map have the option of printing it themselves or contacting a Statistics Canada regional office where the printing can be done for a fee as in the past.
About 6,000 maps have been produced in seven series based on the geographic units of presentation and the geographic units for which they are presented. The series are:
Census Divisions
Economic Regions and Census Divisions
Census Metropolitan Areas and Census Agglomerations
Statistical Area Classification, 2001 Census Subdivisions
Dissemination area maps replace the old enumeration maps series. Their purpose however remains the same - to present the smallest geographic units for which census profile data are available. The presentation in maps of dissemination areas for the small census agglomerations without census tracts is new for the 2001 census. Enumeration areas were not presented in this way previously. This particular series contains 93 maps. Also new for 2001, the Statistical Area Classification, 2001 Census Subdivisions map shows the degree of metropolitan influence on census subdivisions that are not part of either census metropolitan areas or census agglomerations. Intended to provide a better delineation of non-metropolitan Canada, a variety of census data for 2001 will be available by this "metropolitan influenced" categorisation.
Only one of the three enumeration area maps series in 1996 was produced using automated systems. For 2001 all three dissemination area map series have been produced this way.
Boundary and Road Files
Digital boundary files are available for the new dissemination areas as well as all other larger units of the census geography hierarchy. Boundaries include shorelines as well as other selected water features. Road network files serve as a reference layer in conjunction with the boundaries and now cover all of Canada. Names of the roads are included on the files. Files can also be provided for smaller areas as required. A skeletal version of the road file is also available for applications that do not require the detail of the complete file. Boundary and road files come from the new geographic base and are vertically integrated.
GeoSearchÔ
GeoSearch is a new Internet tool that is being introduced for the 2001 Census. GeoSearch provides a fast and efficient way to get census information on places of interest. It is meant for the general public and will answer their typical question, "What’s the population where I live?" The tool lets the user select their area of interest, from small, sub-neighbourhood units up to provinces and Canada.
GeoSearch users find places of interest through place name searches or by working with the map presented on the screen. Query results will indicate:
Before the 2001 Census and the introduction of GeoSearch, to get the same information now available from this tool, a person would have to go to a paper publication and, if they knew the name or census code number of a place, search the tables until they found it. If they didn’t know the census code number, they would have to obtain a census reference map, find the area in which they were interested, note the code number and then look up the area in a publication.
Built using Esri Arc/IMSâ , GeoSearch is made possible by the new national digital holdings and the development of a geography data dissemination infrastructure that serves information for requests via a ColdFusionâ interface.
Statistics Canada’s Internet Site
The census geography pages on the Statistics Canada Internet site bring together many of the geography products and tools and served as the primary source of information for the first releases of the 2001 Census in March 2002. As noted already, the Reference maps and GeoSearch are major components of the site. The site also contains data tables of census population and dwelling counts for geographic units ranging in size from municipalities to provinces. The ColdFusionâ database that serves these tables allows the users to sort the table by column in ascending or descending order.
Postal Code Products
A number of postal code products have been produced for past censuses and this continues for 2001. Derived from information gathered on the census questionnaire, items such as the forward sortation area cartographic boundary file and the postal code conversion file, are tools that increase the utility of census data.
In the Future
Geo-coding services that support census data tabulations for user defined areas will be available in the summer of 2002. The introduction of the block as the basic geographic unit will lead to major improvements in the quality of geo-coding for non- metropolitan areas. Until now, geocoding for these areas used enumeration areas as the unit of aggregation. The size of rural enumeration areas was often very large and this limited the degree of precision for data retrievals. For the first time, users will be able to have custom data tabulations prepared for all areas of the country from a block level base. The result will be relatively consistent data quality from custom retrievals nationally as well as significantly improved data precision in the more sparsely settled areas of the country. In large urban centres, custom retrievals will be available from aggregations of block face data.
The spatial data infrastructure will be updated regularly with the addition of new roads, boundary changes and other features. These benefits will be passed on to data users. Revised census boundaries and an updated road file are planned for release on a regular basis to support data analysis and mapping.
Major benefits for the collection of the census have been realised through the implementation of the new geographic infrastructure. The benefits that the new infrastructure will deliver to data users through the dissemination of the 2001 census products are also starting to appear and this should continue as we move towards the next census in 2006.
References
Geography Staff, Statistics Canada, 1980. Geography of the 1981 Census of Canada, Ottawa.
Methodology and Systems Branch, Statistics Canada, 1972. Geographically Referenced Data Storage and Retrieval System – A Technical Description, Ottawa.
Parenteau, R. and D Wall, 1997. "Geographic Data Base Updates and Local Inputs" presented at Urban and Regional Information Systems Conference, Toronto.
Tallon, P., 1997. "The Implementation of GIS in the Census Dissemination Mapping Program of Statistics Canada", Working Paper Number 21, Conference of European Statisticians, Work Session on Geographic Information Systems, Brighton, England.