Democracy meets GIS
Wednesday, 08 August, 2012
Good maps, bad maps These new competitors have led traditional GIS suppliers to investigate scalable and cost effective ways of delivering services, including cloud computing. For instance, at the United States Environmental Protection Agency and Department of Agriculture Forestry Service, Vining says ESRI has been hired to assemble a portal of various data sets hosted on Amazon Web Services and publish these to the general public. Similar projects are afoot closer to home. In South East Queensland, a group of councils is banding together to create the One Source One Queensland project, which aims to make all underground infrastructure locations accessible through a single cloud data store and displayed using Google Earth. These consumer tools are evolving quickly also. Google in particular is working hard to claim new territory, with 1 billion copies of Google Earth downloaded. Its Google Map Maker tool, available since April 2011, crowdsources the addition of new landmarks and information to the map – and was only recently launched in Australia. Dylan Lorimer, product manager for Google’s Maps Engine, says numerous US states have used its tools to create their own maps and globes. In the State of Alabama, for example, the project started by publishing data from one county. “Anyone could log in and look at all of the accurate current relevant data for this county,” Lorimer says. “Over the course of a year, a complete state-wide view was built where each of the counties retained the ability to update the data at any given time.” While Google may be the hero of the desktop cartographer, it is also inadvertently the biggest culprit in terms of geographic inaccuracies. According to the chief information officer at the Victorian Department of Sustainability and the Environment, Bruce Thompson, Google Maps uses data from non-authoritative sources of address information. For example, data coming from the Telstra’s advertising division Sensis may be sourced from Telstra's own infrastructure rollout – but in the case of new subdivisions this data may be logged before construction is completed, and contain lot numbers rather than residential addresses. Hence lot numbers are confused as addresses in Google Maps. Thompson knows of one street in Kew, in inner-east Melbourne, which has 187 dwellings that are addressed as ‘Lot 1’. “The problem there is they have a singular indifference to data quality, although they are starting to come around,” he says. Google for its part pledges to provide to the most up-to-date maps possible, and also encourages users to report errors. This problem has been felt acutely by NBN Co, which has relied heavily on geographic data to plan its nationwide fibre rollout to what will eventually be more than 10 million premises. In May, chief executive Mike Quigley told a Senate estimates hearing that as much as 30 per cent of data for multi-dwelling units was not consistent.
Scaling for clean data As Thompson explains it, while spatial classification systems such as GPS coordinates are strongly defined, address (cadastral) and administrative boundaries can vary remarkably. The digitisation of information from paper maps took place in the 1980s and 1990s, but those maps often varied greatly in scale. Hence some boundaries in North West Queensland are kilometres out of alignment. “Spatial is one of the fundamental indices of our environment,” Thompson says. “But we still let people make up their own addresses. “You wouldn’t allow that to happen in the Internet space with IP addresses.” Thompson is working in conjunction with numerous parties to create a unified, standardised marketplace for spatial information that will reduce data conflict. Also involved are the research organisation CRCSI, data supplier PSMA Australia, the ANZLIC Spatial Information Council and the Spatial Industry Business Association. Their goal is to create a federated marketplace that spans state and national boundaries across Australia and New Zealand. “You won’t have to have a four year degree in geospatial science to utilise these systems. You can be an ordinary citizen and get access to them.” “We want the marketplace to be something where there are hundreds of thousands of resources, be they data sets or products or services or processes,” Thompson says. The marketplace would make previously undiscoverable spatial datasets available to anyone who wanted them, with an appropriate remuneration mechanism for data owners. A demonstrator project involving 1100 datasets will be evaluated over the next six months, but full implementation is at least a year away. While commercial alternatives also exist or are being proposed, Thompson says these remain proprietary and cannot be federated. The marketplace is a key initiative for the CRCSI. CEO Peter Woodgate says it faces a significant challenge in adhering to the ease-of-use expectations created by Google and Bing. “You won’t have to have a four year degree in geospatial science to utilise these systems,” he says. “You can be an ordinary citizen and get access to them.” Woodgate is also wrestling with another ramification of the democratisation of GIS: the desire for consumers to add their own content. “Verifying the quality of information is a critical aspect of the work that will need to take place over the coming year or two in order to ensure that we do have some quality control filtering mechanism,” he says.
Virtually yours While the marketplace is an audacious project, it pales in comparison to the mission of Virtual Australian & New Zealand Initiative (VANZI), which aims to recreate the entire geography of both countries in virtual form. VANZI chief executive officer Michael Haines says the idea originated from his background in logistics and realisation that infrastructure planning could be improved if projects were modelled and trialled virtually. An initial presentation at the spatial@gov conference in 2010 garnered support from the CRCSI, the Municipal Association of Victoria, the Victorian Partnership for Advanced Computing, and NICTA. VANZI has identified a range of key technologies including GIS, building information modelling (BIM) tools and precinct modelling tools needed to represent complete environments virtually in 3D. “It is those technologies together that create the virtual world, but in order to create it you need data,” Haines says. But even if VANZI gains the necessary support, the project itself may take decades to unfold. “It will take us 20 years to build a virtual world that has taken us 200 years to build physically,” Haines says. “It will take us 20 years to build a virtual world that has taken us 200 years to build physically." All the while, spatial data volumes are growing rapidly. The advent of robotic drones, for instance, is making the gathering of aerial data cheaper, and encouraging its collection on a more frequent basis. Other initiatives, such as NICTA’s AutoMap, are creating new ways for efficiently and accurately gathering ground-based data. AutoMap has worked with the City of Greater Geelong to inventory road signs across 2500km of urban and rural roads. It uses the vehicle-mounted Earthmine system from Geomatic Technologies to collect 3D panoramic imagery, and then analysed this to identify and locate more than 20,000 signs. According to NICTA commercialisation manager Paul Stapleton, the long-term goal is to attach this technology to fleets of vehicles that will constantly update roadside infrastructure data for use in mapping. Further initiatives, such as various environmental sensor deployments being trialled by the CSIRO in the Murray Darling Basin, promise to increase the volumes of live data made available. Ultimately, CRCSI’s Woodgate says GSI consumers may experience the same real-time access to data they expect through common web searching for news or shopping. This serves to strengthen the case for the proposed spatial marketplace. “The more current the information," he explains, "the more valuable it becomes to an end user, so the more likely people are to want to use the marketplace and grow the marketplace."
Data streams and city dreams: how real-time data is transforming urban living
We are entering the era of smart cities, where cutting-edge technologies like real-time data...
Data challenges in city infrastructure and transport systems
In today's landscape of digital transformation efforts by cities and transportation...
Automated number plate recognition is essential for law enforcement
Automated registration plate recognition can detect number plates on vehicles and, by using...