Better Practice Checklist - 17. Spatial Data on the Internet

May 2004 (organisational details updated January 2008)

Introduction

Australian Government departments and agencies are encouraged to maximise the potential of the Internet for information provision, administration and service delivery. The use of spatial information (data that has been connected with a geographic location) is a powerful way in which the delivery of a range of information and services has been enhanced. A key role of the Australian Government Information Management Office (AGIMO), Department of Finance and Deregulation is to identify and promote 'Better Practice'. This checklist has been created to help agencies to maximise their use of new technologies, by giving some guidance on the management and provision of spatial information online and of links to appropriate resources offered by key agencies.

This checklist suggests that a number of issues should be considered when providing spatial data online. The items in the checklist are, however, not mandatory. The checklist has been provided as a guide to help agencies to consider the key issues.

This checklist was created for staff responsible for websites, as well as IT managers and other managers responsible for spatial data. This checklist focuses on non-technical issues.

It should be noted that the checklist is not intended to be comprehensive. Rather, it highlights key issues for agencies. The checklist is iterative and draws on the expertise and experience of practitioners. The subject matter and issues are reviewed and updated to reflect developments.

Download PDF of Checklist 17 - Spatial Data on the Internet [ - 301 KB]

Acknowledgments

This checklist was developed with the assistance of Australian Government agencies. In particular, we would like to thank the Office of Spatial Data Management www.osdm.gov.au [].

What is spatial data used for?

People use spatial data every day to provide answers to questions and to aid decision making. This can range from simple questions like How can I get to a particular shop in the next suburb? to complex ones like How can I predict the potential impact of climate change on the distribution of the Mountain Pygmy Possum? Spatial data can be very effective in supporting a vast array of decisions by governments and the wider community.

Governments hold, or have access to, massive stores of spatial data. Increasingly, they are making this data accessible to others through spatial data products and services. Not only are they seeking to make basic data more available, but they are also packaging data in a wide variety of ways to assist decision making in both the public and private sectors.

Delivering spatial data online involves three distinct areas for consideration:

the spatial data
the information technology
the output form.

While these are a standard list of considerations, there are a few issues specific to the spatial nature of the data. The following explanation of basic technical concepts, and the checklist, are intended to assist developers and managers of spatial data products and services with the main issues. Much of the detail will need to be referred to experts for their advice.

A variety of services use spatial data. These include:

Maps. Map services include the provision of maps that can be manipulated online. The user can zoom in and out, recentre the map and use other features. An example is the Australian Natural Resources Atlas http://audit.ea.gov.au/anra/atlas_home.cfm [].
Routing information. This is information that gives the user directions about how to get from point A to point B, which may or may not include a map (for example, www.toiletmap.gov.au []).
Geocoding / reverse geocoding. This is the generation of a coordinate given an address or, alternatively, of an address given a coordinate (for example, www.whereis.com.au []).
Web services. These provide access to data, which can be distributed across multiple agencies so that other applications can use this data for their own purposes. In the case of spatial data, web services may provide data for presentation in a map. The NSW Natural Resources Atlas www.canri.nsw.gov.au [] relies heavily on web services to provide the spatial data requested.

Summary of Checkpoints

Some basic technical concepts

Check box Determine what spatial data is and whether the agency has any

Check box Consider whether the agency wants to make its data available spatially

Check box Consider the difference between vector and raster data

Check box Understand the Australian Spatial Data Infrastructure (ASDI)

Planning

Check box Plan the product or service to be provided

Check box Consider the product's lifecycle

Check box Consider the needs of potential users

Check box Consider appropriate software

Check box Consider appropriate cartography

Data issues

Check box Plan data requirements

Check box Identify the base data that best suits the application

Check box Ensure that the data is fit for the purpose

Check box Negotiate necessary partnerships

Standards

Check box Apply relevant technical standards

Check box Assign appropriate metadata

Checkpoints

Some basic technical concepts

Determine what spatial data is and whether the agency has any

Spatial data is data that describes the attributes of some object or thing occurring at one or more locations or in a region in geographic space. The object could be an office building, water reticulation network, mangrove forest, calving grounds for Humpback Whales, or suburbs that have more than 50% of households with incomes greater than $40 000 per year. Data is normally managed as part of some theme: built environment, utility infrastructure, vegetation, population dynamics, demography and many others.

Spatial data is made by connecting data about a theme with a spatial model of its geographic distribution. For example, the simplest method is to connect the latitude/longitude coordinates of the location with each observation in a dataset. Similarly, a spatial model of electorates can be made and then linked with data about any theme for which information can be collected by electorate (for example, median income, export revenue, unemployment, percentage agricultural land).

Agencies that have collected location information, whether intentionally or otherwise, could consider making this data available in some spatial form, if appropriate. Location information may include a street address or the name of a facility such as a hospital. Even just a town name may be suitable for some applications, where a link to a gazetteer can be used to supply a geographical reference.

Consider whether the agency wants to make its data available spatially

The first reason for making data available spatially is to make maps. A map is a graphic representation of a dataset, just like a graph. Like a graph, a map can quickly communicate the salient information from a large amount of data.

Secondly, a spatial dataset can be used to highlight the distribution of themes across any region of the country. A spatial attribute also allows ready connection, based on location, with other data that may have been collected by others for unrelated purposes. This provides the ability to analyse data in new ways. For example, a map of patients diagnosed with a rare but debilitating disease can be matched with a map of factories using or producing a particular chemical, thus suggesting a causal link that can be further investigated.

Consider the difference between vector and raster data

Vector and raster are different ways of storing data in a digital file. Attributes about themes are usually observed/collected/modelled spatially as changing in either a discrete manner (vector data) or a continuous manner (raster data) over geographic space.

Vector data. Discrete spatial units include buildings or power poles (points), roads or rivers (lines), or suburbs or electorates (polygons). These are represented as vector data, where everything with the same representation, including symbol, colour and shading, has the same attribute or the same value for the attribute being represented. It follows that everything not having that representation does not have that attribute or has some other value for that attribute.
Raster data. Continuous spatial units do not have clear boundaries. Rather, they grade into one another. Examples include land elevation, rainfall distribution, noise levels around an airport or pollution around a factory. These are more effectively modelled as raster datasets. Rasters are best thought of as a uniform grid covering the geographic space. Each grid cell has a value representing the attribute at that particular location. Unlike with vector data, the attribute values may well vary within the grid cell, but the value is chosen to be 'representative' of the cell in question. If intra-cell variation may cause a problem with any subsequent analysis, the grid size can usually be varied to minimise any distortion. Further, as another inherent feature of continuous data, the attribute values in a cell are usually related to those in the neighbouring cells.

It is possible to have vector datasets representing continuous features (and vice versa). There may be some very good reasons for having data in this form. However, units shown in a particular way may have some internal variation. Whether this proves to be a problem depends on the degree of variation relative to the nature of the question. In some cases there will be no difficulty. In others, the interpretation derived from the spatial product could prove to be significantly flawed.

Understand the Australian Spatial Data Infrastructure (ASDI)

The ASDI is a framework of people, policies and technologies that expedite ready access to, and use of, spatial data resources held around the nation www.anzlic.org.au/infrastructure_ASDI.html []. The ASDI is being facilitated by all Australian Government jurisdictions under the leadership of ANZLIC - The Spatial Information Council www.anzlic.org.au [].

The ASDI supports the spatial data activities of governments and the wider community. From defining the mathematical origin of the Australian spatial domain through to providing access to high-quality base datasets, the ASDI underpins sustainability and triple-bottom-line outcomes through enhanced social, economic and environmental efficiency and productivity.

Planning

Plan the product or service to be provided

The deployment of any online system should be backed by an approved business case that identifies the business drivers for the service. Spatial systems can be very attractive and graphically enticing; but as with other initiatives, the business case needs to justify the investment.

In determining the functionality of the online service, agencies may wish to consider the output that the system will produce for users, the level of access users will have to the data, and the thematic layers of information that will be provided.

Output issues include the choice between textual and graphical formats and the actual content of the output. This includes the core information that the system will return and the supporting information. For graphical output, this would be referred to as the 'cartographic style' of the map. This can be crucial, as maps have a long history of being used (both intentionally and unintentionally) to present information in a less than impartial manner.

Agencies may also consider the online nature of the output. For example, can a particular map be printed or saved? And if so, can it be done at a print quality resolution rather than a screen resolution?

User access issues include whether users will be given access to the raw data or to an interpreted representation like a map, and whether they can see the whole database or just little bits at a time.

Agencies may also consider how users will expect to interact with the data. For example, will a query/response system be appropriate? Will the service start with a picture of the whole of Australia and expect the user to zoom in, zoom out and pan around until they have what they want? Can the user select the layers they actually want to see and turn off the ones they don't want to see?

Thematic layer issues include the number of layers, any association between the layers (for example, same theme at different dates), the data structure and file format. For example, agencies may consider the difference between a web map service and a web feature service:

Web map service. This will limit access to a 'picture' map of the data in the form of an image. This representation is static and the individual components cannot be further queried or analysed.
Web feature service. This will provide a similar graphical output, but as data elements that the client side will render into a map locally. This allows a much greater degree of flexibility for the client to customise the output to their needs. Agencies may choose to offer the ability to connect to or download the entire dataset in its native format for use in an offline application, such as a modelling package.

Consider the product's lifecycle

Consider a maintenance and lifecycle strategy for the service. Issues may include:

Maintaining staff skills in the systems that have been developed, especially where the initial development was outsourced.
Load demand on servers and communication networks, which can increase dramatically if the site becomes popular.
Supply chains for base data and thematic data.
External factors that may render the data in the system obsolete. For example, failure to reflect the gazettal of new electoral boundaries could be embarrassing and cause users to lose confidence in the service.

Agencies may also wish to consider the intended users' business cycle and when the information they provide is most useful to users. This may help to determine the frequency of revisions. Knowing when to review and possibly replace the whole system is also a key decision.

Consider the needs of potential users

When developing an online system, attracting and retaining users may be a significant issue. Consideration should be given to the look and feel of the site, as well as to the usability and the functionality of the interface. A technically elegant solution that is too difficult to use will not attract many users. Similarly, consider ways of minimising the volume of data that needs to be sent down the line to the user's web browser. This is especially important to users with low-bandwidth access. Agencies may consider having their server do as much of the processing as practicable, and send just the image of what the user will actually see. Agencies should also be aware of the needs of sight-impaired and other restricted capacity users, and of the Government's requirement that online government services be accessible for these users.

Further information about developing appropriate services and marketing is available in Better Practice Checklist 20, Marketing E-government.

Consider appropriate software

Starting with a consideration of the agency's existing IT environment (particularly the spatial IT environment if the agency already has current capacity), and the IT strategy, agencies may consider the particular software required.

Delivering a standard spatial dataset may be as easy as deploying a registration form and a download link. A web map service may require a choice between a variety of commercial off-the-shelf and open-source web map servers. A textual query response system for real-time data may require in-house development in order to be tailored to a specific solution.

Spatial software systems usually require configuration, and out-of-the-box systems will require customisation for the specifics of the application and to match the graphical look and feel of the site.

Many vendors offer additional functionality, like a travel-routing server as an add-on to the base product. Each of these additional functionality requirements will add to the work to be done to get the system operational, and potentially to the load on the servers for additional data processing. For example, some vendors recommend an additional server just to do the routing processing.

Consider appropriate cartography

Cartography is the art of communicating through maps. The same information, presented in different ways, can result in very different perceptions of the information.

Computer screen cartography differs from paper cartography. (See http://kartoweb.itc.nl/webcartography/webbook/index.htm [] for a full discussion of web cartography.) The most important differences are that, in computer cartography, less display space is available and, in most cases, the user is able to change the scale of the map, affecting the amount and resolution of information that can be displayed on one screen. On the other hand, computer cartography is potentially much more powerful through its interactivity, which allows the user to tailor the result to their requirements. These factors will impact upon the choice of datasets for display as well as on how they will be displayed.

Data issues

Plan data requirements

After determining the functionality that is to be provided by the application, consider at an abstract level the spatial data that will be required. Data has the potential to be a very expensive element of a web-mapping application, especially if it has to be collected. Access to external spatial data will often require the negotiation of licence agreements and, in some cases, the payment of a fee for its use in an online system.

Base data. For most graphical applications some base information will be needed: the coastline and state borders for a national map; main roads and location names for a regional map; local roads, localities and perhaps other significant location identifiers for a local maps. Most of this sort of information is available from data suppliers
Thematic data. This is the information that people consider in order to make a decision. This is likely to be the element most valuable to users because it comes from the organisation's activities.

Any thematic data to be provided online will need to have a spatial link added. The type of link will vary, depending on what the thematic data actually represents. Consider whether the data represents:

locations (for example, city, suburb or particular location)
areas (for example, Statistical Local Area, catchment, specific management regions), or
other administrative boundaries.

Also consider whether your agency has a clear spatial definition (even better, a spatial dataset) of these points or boundaries.

Boundaries can change, and agencies may wish to consider strategies to ensure that the boundaries they are using are correct. Agencies may also wish to consider the issue of making temporal comparisons between datasets that have changed boundaries.

The Australian Spatial Data Infrastructure (ASDI) aims to build a coordinated infrastructure of spatial data. Agencies making spatial datasets are encouraged to consider coordinating their approaches with the ASDI www.anzlic.org.au/infrastructure_ASDI.html [].

Identify the base data that best suits the application

Spatial data is generated and stored in a range of formats. Most formats are file-based, although there is significant movement towards the use of spatial capabilities in mainstream corporate database systems. Differences in formats can cause difficulties if data from different sources needs to be used. Many of the file-based data formats are proprietary and may need translating for use in the agency's online application.

The Australian Spatial Data Directory (ASDD) asdd.ga.gov.au/asdd provides access to directory-level information (metadata) about spatial data that is available throughout Australia and New Zealand. Also consider other information that may be held by local government and the private sector.

Ensure that the data is fit for the purpose

The quality of the services provided will depend on the quality of the data used. Consider:

Appropriate data resolution. If an application is to enable users to view data over small areas (such as a block of houses), it is usually not appropriate or useful to display datasets meant for displaying large areas (such as a state). Similarly, it may not be helpful to use datasets captured for use over small areas if they will be viewed only in the context of larger areas. In such cases, the extra information stored in the high-resolution dataset will take a long time to download to the user and clutter up the resulting display, making it more difficult to understand.
Data currency. A number of spatial datasets are dynamic in nature. That is, the information contained within the dataset changes over time. Consider setting up measures to ensure that datasets delivered on a website remain up-to-date. These can be as simple as assessing the 'shelf life' of each dataset and setting up calendar reminders to check whether new versions have become available.

Negotiate necessary partnerships

As agencies may need to source their data from other providers, they may wish to consider:

Which agencies hold data or other resources that are crucial to the service?
Have all the arrangements for data access and ongoing support been negotiated and confirmed appropriately?
Where are the risks, and what are the contingency plans, if a key supplier is no longer able or willing to maintain its data to the standard required or to make its data available under established arrangements?

Standards

Apply relevant technical standards

The standards environment for spatial information technology is currently very dynamic. The emergence of a suite of abstract standards from the OpenGIS Consortium (OGC) www.opengeospatial.org [] and complementary implementation standards from ISO TC211 www.isotc211.org [] is leading the open source push in the spatial information technology arena.

Increasingly, technology vendors are offering products that meet some or all of the promulgated standards. Agencies may wish to check to ensure that the applications they are implementing meet current standards. Most proprietary systems are capable of being used in the development of an online system and appear similar to the standards, but they may lack interoperability for more advanced applications.

Assign appropriate metadata

Metadata is data describing data or information. Spatial metadata is information that describes spatial datasets. It facilitates the storage and retrieval of information about particular datasets through the Australian Spatial Data Directory (ASDD). Standardisation of the metadata model facilitates data exchange within and between organisations.

Metadata created for spatial datasets should conform to the ANZLIC Metadata Guidelines. These guidelines encompass the Australian Government Locator Service (AGLS) metadata standard. Further information about the ANZLIC Metadata Guidelines is available at www.anzlic.org.au/infrastructure_metadata.html [].

For further information about AGLS metadata for other online resources, see Better Practice Checklist 6, Use of Metadata for Web Resources.

Other Better Practice Checklists

Download PDF of Checklist 17 - Spatial Data on the Internet [ - 301 KB]

Contact for information on this page: AGIMO Better Practice Team

Skip links

Main site links

Contact and help

Menu

Better Practice Checklist - 17. Spatial Data on the Internet

Introduction

Acknowledgments

What is spatial data used for?

Summary of Checkpoints

Some basic technical concepts

Planning

Data issues

Standards

Checkpoints

Some basic technical concepts

Planning

Data issues

Standards

Other Better Practice Checklists