Prof Arup Dasgupta
While the importance of the expert group of surveyors and cartographers in the context of geospatial information cannot be devalued, amateurs are becoming increasingly active and the sharp boundary between the expert and amateur is getting blurred. Prof. Arup Dasgupta, Managing Editor, analyses the process of democratisation of geo-information
The public at large are the biggest stakeholders in the field of geographical applications. The daily acts of transportation, location and recreation, to name a few, are all geographically referenced. In fact, it would not be wrong to say that geography provides the leitmotif to human existence. An individual is a resource for geographical information that is localised and personal, yet the average citizen hardly had any role to play in the geospatial world. This fact was realised by local level planners and the first attempt to draw on this wealth of local information and include it in the more formal technological world of GIS was termed “public participatory geographic information systems, or PPGIS. The attempt in PPGIS is to quantify local information and knowledge and in the process give ownership of the system to the stakeholders.
With modern systems like GPS receivers, GPSenabled devices like phones and cameras and the interactive Web getting popular, a whole range of geographic data collection and dissemination activities emerged, characterised by the generic term “user generated content,” or UGC. Terms such as volunteered geographic information, neogeography and crowdsourcing are specific instances of UGC. These terms cover the act of the collection and dissemination of geographic information by individuals who may or may not be trained surveyors and geographers but who have sufficient knowledge to be able to observe, record and report a geographical feature or event. VGI is characterised by a formal set up where the public can contribute; a typical example being Google Map Maker where tools are provided to the volunteer to enter data which is further subject to review and moderation. Wikimapia is another example but here the review process itself is driven by volunteers. Neogeography, on the other hand, lacks a formal structure. It is more intuitive, personal and perhaps even idiosyncratic but it is an expression of an individual geographic view – like a mashup.
A related term is crowdsourcing where each individual is considered as a sensor and thus is a source of geographic information. Crowdsourcing can be solicited or involuntary. A radio station which looks for traffic reports from citizens is an example of a solicited crowdsourcing. Studying the movement of mobile phone locations to determine traffic congestion is an example of involuntary crowd sourcing. The latter type of crowd sourcing raises many issues of privacy if it is used to identify individuals and tracks them to determine their preferred locations and personal interests.
THE SOCIAL CONTEXT
Public participation in geospatial activities is a part of the continuum of public participation in ICT. Web 2.0 has given rise to the emergence of Web-based communities which focus on social networking, social book marking, blogging and Wikis. VGI and neogeography arise from the desire of the average person to tag data to specific locations of their interest and share this data with others in the same way as they share on say, Facebook. Information has become pervasive and ubiquitous and geographical information is following this trend, resulting in a fundamental shift in geographic knowledge production and practice. Chris Thomas, Government Market Manager, Esri, reflects this when he states that "Esri saw the early beginnings through citizen and volunteer group involvement during disasters where disparate groups began to report road conditions, need for water, lack of electricity, etc. The use of GIS through light weight applications on smart devices began to increase through focused citizen engagement activities. We witnessed groups and individuals using GIS to support a wide range of activities today, from reporting graffiti and pot holes, to relaying information on the location of endangered species, supporting real time collection of broadband connectivity for future infrastructure planning, supporting the inventorying of trees in an urban forest, and much more. The list of activities grows each and every day."
While the importance of the expert group of surveyors and cartographers in the context of geospatial information is not to be devalued, at the same time amateurs are becoming increasingly active and the sharp boundary between the expert and amateur is getting blurred. The advantage that the amateurs have over the experts is that their information is up-to-date and localised but the disadvantage is that the information is not as accurate and reliable as the information from an expert. This process of democratisation, therefore, is not without its dangers. As in other social media, a lot of information can be misleading if not downright erroneous.
The Report of the EuroSDR Workshop on Crowd Sourcing for the Updating of National Databases held in 2009 identifies several types of volunteers.
"Groupies" are small groups of map (or any geodata) lovers which produce trustable and very valuable data of great value.
The "casual users" are hikers, bikers, mountaineers etc. This group partly overlaps with the groupies but are distinguished by a lesser effort and therefore less valuable data production.
The "experts" are active people and leading map experts in organisations like mountain rescue, fire brigades, civil protection, traffic guides, etc. The experts have to be identified by, and committed to, the mapping agencies with (financial) agreements.
The "media mappers" are potentially large groups, activated sporadically by regional up to international media campaigns (television, radio, internet forums, print media, etc.). These are mostly once-off mappers specially motivated by competitions, mapping parties, etc.
The "open mappers" are small groups which spend a lot of time to contribute very valuable and large information to open source data sets or data systems (OpenStreetMap, MapShare, Google MapMaker, etc.).
The "passive mappers" are owners of mobile phones which incorporate GPS positioning. The use of this information is restricted by law and/or acceptance of the individuals.
The "mechanical turks": The Amazon Mechanical Turk is a crowd sourcing marketplace and service where human beings can contribute to posed tasks for a monetary payment. They can be used to gather needed data for mapping agencies, but the delivered data would need careful checking.
THE ISSUE OF AUTHENTICITY
As can be seen from the profile of volunteers (Box 1) the reliability and authenticity of the VGI ranges from 'low' for data from casual users to 'high' for data from experts. The acceptance of datasets from such a variegated set of volunteers requires special effort. In general, quality of data has multiple dimensions such as absolute and relative geometric accuracy, currency, topological correctness, the accuracy of the metadata, the legal aspects and the validation stamp by an official agency. VGI may not meet all these parameters. The main issue is the fitness of data for the purpose intended. VGI data to be used in a disaster situation should be up-to-date and usable for navigation even if the accuracy is not of map quality. On the other hand, data which has a higher persistence level like a road has to be accurate and certified by an authority.
For example, MapmyIndia, an LBS portal, ensures authenticity by validating such crowdsourced information with their large in-house team of 400+ professional field surveyors and 150+ map specialist engineers. Hence, MapmyIndia data is crowd-augmented yet professionally created and verified, creating a continuous cycle of best, freshest and most accurate data available for India, says Rohan Verma, Director, MapmyIndia.
THE NMO VIEW
Ng Siau Yong, Director, GeoSpatial Division, Singapore Land Authority feels that while the authoritative data has to necessarily come from the national mapping organisations (NMOs), current data is where volunteers can contribute. According to Yong, "There is a sea of potential for VGI. People are now equipped with technology and this empowers them to provide insight to things happening at various locations. The ability to report an incident speedily is very common these days as seen through Twitter and Facebook. OneMap has recently launched a set of crowdsourcing tools which will enable users to collect crowdsourced data through a plug-in which one can embed on their webpage. A set of crowdsourcing tools or APIs on OneMap can be used by developers to create their own crowdsourcing tools on their websites or create a mobile application to capture data on the move."
Georg Gartner, President, International Cartographic Association (ICA) and William Cartwright, Department of Land Information, RMIT University, Australia both feel that from the perspective of cartography, VGI is a very positive development. “Open Street Maps has made an impact on NMOs. The task that NMOs have done in the past as an authority will partly be covered by both commercial and volunteered communities. However, a cartographic product is a standardised product like a newspaper or a movie, but a VGI product is a customisation of that standardised product. Issues of accuracy and reliability of CGI data will decide on its acceptability,” they say.
These views are reflected in the conclusion of the Report of the EuroSDR Workshop on Crowd Sourcing for the Updating of National Databases held in 2009. It states that "a strong collaboration of the mapping agencies and the research communities in the field of user generated content, which is identified as an important and valuable input in map or geodata updating procedures, content generation and interaction with the users. There is a lack on experiences, knowledge on interaction schemes, on techniques, on legal aspects, on production process integration aspects, etc. The dialogue between crowds and mapping agencies is in most cases not established or is at a very basic level. How to establish, support and motivate a community to achieve tailored user generated content is seen as one main field of investigation."
The not so friendly term of ‘crowdsourcing’ now has been in use for quite some time, as well as the rather anonymous-sounding ‘volunteered geographic information’. With ‘citizens as sensors’ it gets much more specific and personal, immediately raising concerns of privacy and perhaps bringing up a sense of responsibility as ‘citizens’.
Some simple categories help with getting categories straight: we can volunteer spatial data (like in OpenStreetMap), or data about ourselves (by checking into a geosocial service point-of-interest). We are aware of our actions being recorded as a contribution to a community –like in the cases mentioned above- or not, which actually would mean the generation of ‘involuntary geographic information’. Again, the latter can be broken down into anonymously tracking a ‘crowd’ (mobile phones checked into a cell, or handed over between cells – e.g. when monitoring traffic), or the recording of individually identified traces, e.g. through personal customer records or payment histories.
From a geospatial data point of view some or all of these ‘people as sensor’ recordings obviously are considered highly attractive for a multitude of applications domains, many of them novel. Social sciences and many of their applied frameworks like geomarketing, security in public spaces, mobility management and urban design are more than keen to leverage the potential of sensing what’s going on in social spaces.
Without entering the complexities of the privacy debate and discussing the concept of ownership in one’s own data, it is clear that location adds an entirely new dimension to privacy. As much as presence in a public space might seem to be all but private, we now recognize from many practical examples that a person’s location is an extremely powerful key allowing the linking of otherwise independent data sets. The ‘spatial join’ of personal records is much more critical than the simple tracking of an individual, raising legitimate concerns about where the ‘people as sensors’ evolution is going.
People-as-sensors, or people-being-observed might be the most important distinction to be made, but unfortunately this often is kept in a fuzzy state. Thus only ‘full disclosure’ of any sensory role creates legitimacy, as does an explicit opt-in approach to any kind of identifiable tracing – keeping in mind that with ‘spatial joins’, identification is never far away. Whenever we act as citizens, though, e.g. with a desire for public participation, individual identification typically would be intended.
Another facet of ‘people as sensors’ frequently gets caught up in this debate: whenever truly ‘crowd sensing’ is done to observe clustering, movements and collective actions of larger numbers of people, benefits might clearly outweigh the critical aspects. Lots of people indicate collective behavior, lots of cars indicate traffic, lots of customers let us measure business – as long as we stay away from Id’ing individuals, the ‘crowd’ serves as the protective cover of privacy concerns. ‘Crowd’, all of a sudden, thus becomes a much more friendly term!
Prof. Josef Strobl
Director, Centre for Geoinformatics, University of Salzburg, Austria
The main issue with VGI is the lack of standards. However, there is a limited awareness of the standards as volunteers are generally not involved in developing them. As a result, there is limited understanding of their context or utility and therefore very few volunteers "buy in" to the standards. It does not help that standards are difficult to read. Another difficulty is that data must have associated metadata. Most volunteers do not know about metadata and those who know, do not appreciate its value. In fact, data collected by experts often lacks adequate metadata, therefore it is no surprise that VGI has even less metadata. One way of imposing standards without loading the volunteer is to create a gateway through which the volunteered data has to be submitted. Such a gateway inserts the requisite standardisation. Examples of such gateways are Google Map Maker, Open Street Maps, Wikiloc, 360.org, Wikimapia and Eye on Earth.
The Open Geospatial Consortium has addressed the issue of standardisation in VGI and suggests the mapping of its Sensor Web Enablement standards to VGI. VGI could be treated as a complement to remote sensing. Further, since VGI uses Web 2.0, features would be retrieved from the Web through a set of well-defined rules. This is similar to the gateway approach described earlier. Another OGC effort, Open GeoSMS is an open-coordinate short message service (SMS) standard to allow transmission of map information and communications among different platforms of digital maps. The goal is to share location information across operating systems and applications. This standard is user generated and is in the process of becoming a standard.
VGI is currently a 'free for all' movement. While such altruism is to be appreciated, it may fly in the face of legal statutes and restriction in the long run, particularly if and when it starts getting integrated into the authoritative system. Kevin Pomfret, Executive Director, Centre for Spacial Law and Policy, feels that it will be at this point when "lawyers are likely to begin raising the questions like who owns and is responsible for the quality of the data that is collected – the collector, the entity that aggregates the collected data or the 'owner' of the information that is collected. Should the volunteer be solely responsible for complying with all privacy or national security laws or should the organisation that encourages the collection or aggregates the information also be liable? Is it ethical to encourage volunteers to collect and share such information without providing them knowledge of legal risks? If VGI providers do not have adequate answers and/or solutions to these questions, such that lawyers understand the risk and believe it is properly allocated given the business relationship, there could be significant push back on integrating VGI with other data solutions".
Training and documentation are proposed by Kevin to overcome these issues. "Organisations that encourage or solicit VGI should provide adequate training and materials on potential legal risks. The training and the materials should be tailored for the particular region. For example, the legal issues associated with VGI collection in the United States are different than those in China. Organisations can include language in their documents (contracts, license agreements, terms and conditions) that outline such matters as ownership rights in the data, compliance with local laws, steps taken to insure data quality. While such steps may seem premature in the early stages of collection, they will prove very helpful later on as in many cases the strength of the VGI database will only be as strong as its foundation."
IMPACT ON INDUSTRY
VGI presents a very cost-effective solution for data in data-hungry industries like location-based services. For example, MapmyIndia augments its exhaustive geo-database of India with information that comes from hundreds of thousands individual and corporate users of MapmyIndia data and products. Individual users of MapmyIndia's web, mobile and in-car products can easily provide structured and unstructured feedback, such as: adding a new place, editing an existing place, reporting an error, suggesting better routes between places etc. Corporate users continuously provide feeds to MapmyIndia about new outlets which have opened or those that have been shut, as well as changes in routes being used in their daily business operations dependent on MapmyIndia GIS. VGI has a significant role to play in the industry, and must be harnessed effectively to create a strong positive impact. There are examples of VGI which are fully unstructured or unverified which in the end create inaccuracies and problems for users. So, the industry must continuously invest in a strong underlying professional data gathering and verification backbone and infrastructure, observes Rohan.
According to Chris of Esri, "We are witnessing a broadened definition of enterprise GIS. One that includes society, elected officials, knowledge workers, and government workers that may not have viewed themselves a professional that would benefit from GIS or data collected through GIS. Each week we see the emergence of applications we had not thought of coming in from NGOs, start-ups, academia, traditional business partners, and of course, our user community". Esri perceives a real opportunity in leveraging the VGI movement by adapting their products to embrace citizen-to-government applications. Esri's new solutions initiative ArcGIS for Local Government provides templates such crowdsourcing, citizen requests, emergency management data collection for VGI that can be combined with other templates for a full service experience.
Companies such as City Sourced have emerged as data collectors of citizen requests. By linking to a Web GIS like ArcGIS Online and with a better understanding of the workflows, GIS could support such a data source with a product that could support the full operation. The data collected by the citizen could be pulled into the back office where GIS analysis could be performed, executives could view the information on their dashboards, and the information could be consumed into CRM and 311 applications.
IMPACT ON GOVERNANCE
Perhaps the biggest impact that VGI has is on crisis management. It is a well known fact that during the Haiti earthquake rescue and rehabilitation efforts, the most up-to-date data came not from the UN or the military but from Open Street Maps. Volunteer efforts like Ushahidi provided the platform for collecting VGI. In the Haiti earthquake, 25,186 SMS messages and numerous e-mails, Web, and social media communications resulted in 3,596 georeferenced reports that were displayed cartographically via the web and acted on by government and intergovernmental humanitarian actors.
Gartner and Cartwright are of the view that "VGI data might be more useful in urban areas and might replace other activities at either commercial or at the national authority level. This holds a big effect for the future of cartography. The first step of cartography is to get geo data in order to make maps and cartographic products. The cartographic communication process is influenced by users’ needs." This implies that the volunteer can become a major influence in the urban planning and management process. Municipalities do depend on volunteer information for maintenance of their assets. The impact of urban planning could be judged in advance through a process of consultation with and data collection from volunteers. Chris Thomas opines that "in government, VGI is creating a whole new dialogue between citizens and government. The adaptation of VGI is providing a voice of citizens who may not normally have ever engaged with their federal, state or local governments. The data collected by these groups at times may have taken years to collect or may never have been collected at all. The citizen engagement component is bringing GIS to a much wider audience. Not just the people collecting the data in the field, but elected officials recognise the importance of leveraging the communities within their communities. These elected officials have always known what GIS was but now they can realise a whole new use that impacts them in more personal ways. New geospatial workflows are being created. Thus, new analytical processes and visualisation products are coming to the forefront."
The city of Norrköping in Sweden is empowering its citizens to plan the future city centre. The city planning office of Norrköping has initiated a project to communicate plans for the new city centre as well as engage the citizens to contribute with ideas.
The service in Norrköping is based on Agency9s CityPlanner, a Web-based service for collaboration and communication in urban development and infrastructure projects. Agency CityPlanner is provided as a cloud service and allows projects to:
In the service, the user easily navigates in an interactive 3D view of Norrköping and explores the city. Information points provide details in text, images and 3D models. The user can comment and read others' comments directly in the 3D environment. Unlike traditional physical models in closed exhibition halls, the Web services are available to citizens 24 hours a day. This is a new channel to attract the younger generations and people too busy to visit exhibition meetings and hearing meetings, to communicate projects and build positive support for proposed plans. The service was released by Norrköping Municipality in connection with a local real estate event on April 13-15.
"We are overwhelmed by the positive response from citizens and media. Within a few days of the release of the service, there were many qualified comments from citizens proving its value. The attractive near reality presentation of the city in 3D combined with the simplicity in accessing information brings clarity and makes it easy to create interest from citizens. We will continuously monitor the inputs as contribution in our development of the city," says Dag Johansson City Architect of Norrköping.
VGI is here to stay. A connected community will find expression through many means. Social media is a manifestation of this power. Add to this a geo-equipped community and we have geo-social networks, a powerful lobby of volunteers and activists who can add to geographic information. OSM and Ushahidi are examples of how this power can be fruitfully harnessed. Governments and NMOs in particular are aware of the potential of this network. It will take some time before this potential can be harnessed as an authoritative data source. Chris Thomas sums this up succinctly when he says "The world is adapting and communicating quite differently. The road ahead is stimulated by the imagination of individuals that have picked up GIS for the first time and reintroduced new apps at every turn. I for one love to wake up each day to see new things on the horizon."