Draft paper for Environment and Planning A theme issue: Situating Neogeography, August 2011.
Mapping Experience: Crowdsourced Cartography
Martin Dodge Department of Geography, University of Manchester Rob Kitchin NIRSA, National University of Ireland, Maynooth
Abstract This paper considers the emerging phenomena of crowdsourced cartography in relation to ideas about the organisation of contemporary knowledge production in capitalist societies. Taking a philosophical perspective that views mapping as a processual, creative, productive act, constructed through citational, embodied and contextual experiences, we examine how we might profitably analyse collaborative crowdsourced projects like OpenStreetMap to better understand geographic knowledge production in a shifting political economy and socio-technical landscape. We begin by characterising crowdsourcing practices in the wider context of Web 2.0, which some commentators assert is rapidly becoming a new, dominant mode of knowledge production. We then contextualise Web 2.0 knowledge production, drawing upon the ideas of sociologist George Ritzer, and his notion of ‘prosumption’, geographer Michael Goodchild’s idea of volunteerist ‘citizen scientists’, and economic commentator Nicholas Carr’s critique of the ‘ignorance of crowds’. We then go on to discuss the changing nature of cartography in the Web 2.0 era with respect to authorship, ontology, representation and temporality.
Key words: cartography, crowdsourcing, Internet, Web 2.0, authorship, ontology, representation, temporality.
1. Contemporary knowledge production: Web 2.0 A burgeoning set of concepts has been used to characterise the emerging nature of contemporary digital knowledge production in western economies: peer production, free software, Creative Commons, feeds, blogs, cloud computing, social networking, 1
Electronic copy available at: http://ssrn.com/abstract=1921340
mash-ups, torrents, tags, tweets and wikis. These represent new ways to author information and communicate meanings with respect to a raft of increasingly affordable, pervasive and capable digital technologies, such as smartphones, tablets and notebooks. Collectively, they constitute Web 2.0 – a new era of the Internet differentiated by participation as opposed to broadcast and consumption. In Web 1.0 Internet users were empowered by being able to search and access a vast collection of information. In contrast, during the Web 2.0 phase the Internet has become a media through which users can easily and actively participate in the production of the online world, authoring and editing content, and contributing freely to services which share that material with others. In other words, Web 2.0 is a read+write media, in which people add value to sites as they use them (O’Reilly 2005). Such media include social networking sites such as Facebook, blogs, synchronous microblogs such as Twitter, shared gaming spaces and multi-user virtual worlds, mash-ups, file and code sharing services such as YouTube, Flickr and SourceForge, and social tagging and bookmarking.
A radical feature of some elements of Web 2.0 is the use of crowdsourcing to actively design media structures and populate it with content. Crowdsourcing is the collective generation of media, ideas and data undertaken voluntarily by many people. The crowd as a metaphor signifies the power that can emerge from a mass of individuals converging to tackle a set of tasks. In the virtual realm, a crowd can be drawn together across a widely distributed set of actors for little cost in order to tackle very large challenges (for example, mapping whole countries). For individuals or small groups of volunteers this kind of challenge is simply too daunting, but given many thousands of participants the task can be subdivided and solved. Crowdsourcing then is premised on mass participation, with distributed voluntary effort, and a degree of coordination. A key issue is how to beckon the crowd into being and how to marshal it to form a coherent group working to a common goal. Indeed, the coordination aspect is often the most contentious as motivated individuals can be difficult to corral into meeting shared objectives. Many crowdsourced projects that work effectively are based on loose, consensual leadership that seeks to be non-hierarchical and yet also has degrees of egalitarianism.
Electronic copy available at: http://ssrn.com/abstract=1921340
Such crowdsourcing projects challenge traditional notions of knowledge generation by radically transforming the acts of writing, creating, and mapping. In wiki systems, in particular, the authoring of information becomes a collaborative act, so that all participants can contribute to the production of a unified text, able to edit, refine and extend the contributions of others. In opensource software projects many programmers contribute freely their time and labour to produce source code for the benefit of others. In geocrowd mapping projects such OpenStreetMap, people voluntarily collect, clean and upload GPS and attribute data in order to produce surveyed cartographic information. Novel works can also be constructed in the form of mash-ups that use Application Programming Interfaces (APIs) to combine multiple sources of data in ways not prescribed or necessarily anticipated by the providers. By far the most popular type of mash-up is map-based, wherein data from one source, such as photographs in Flickr are fed into another, such as Google Earth, to create a hybrid cartographic service. In all four kinds of knowledge production – wiki, opensource software, geocrowd mapping and mash-ups – data and information are published under a range of new licensing regimes, with many authors working under the open-source rubric, at odds with conventional copyright strictures, and by signing up to Creative Commons licensing that allow others to easily and freely fully exploit digital content for the social good rather than monetary profit.
The apparent willingness of many people to participate for ‘free’ in crowdsourcing projects is undoubtedly based on the fact that they provide genuinely effective platforms to connect socially, communicate meaningfully and contribute collectively. There are, of course, ongoing debates around the worth of these connections and communications, with critics arguing that they are superficial, lacking depth and obligatory commitment (Carr 2007). There are also concerns over the unpredictability of crowds, the quality and consistency of content and metadata created across diversely skilled/motivated individuals, how to provide documented degrees of reliability and generate a sense of trustworthiness, and the extent to which the model is sustainable (will volunteers continue to give ‘free’ labour next week, next month, and next year; what happens when the crowd disperses?).
Perhaps more troubling, many Web 2.0 services are owned by entrepreneurs and newly founded corporate enterprises that seek to produce new models of capital 3
accumulation by unlocking ‘free’ labour and information resources and creating new markets. Indeed, there now exists a dynamic ecosystem of companies offering many different kinds of participatory online services and mobile platforms to support knowledge production in many forms. These businesses gain benefits at three levels. First, they gain directly authored content (posting, uploads, and sharing) and subsidiary authored information such as commenting, tagging, rating, recommending, forwarding, and linking that together can add up to valuable meta-content. Second, they gain secondary financial benefits through advertising revenue. Third, data relating to site activity and content can be combined, classified and ‘mined’ for social patterns and commercially valuable trends and fashions, packaged and sold to third parties. There are clearly two models of crowdsourcing operating with Web 2.0 mapping. The first is commercially owned, such as Google Maps, where the original base information is professionally sourced and additional information and mass checking is derived from users. Access is currently free with income sourced from advertising and cross-subsidisation, but the ontology is fixed. The second is an opensource model wherein the entire mapping platform and underlying cartographic database is open, editable and not owned by a company operating for profit.
The development of Web 2.0 and crowdsourcing raises a number of questions with respect to geographic knowledge production - its quality, reliability and sustainability - and its underlying political economy, as well as opening up fundamental questions with respect to authorship, ontology, representation and temporality. In this paper, we examine these questions, focusing in particular on open crowdsourced mapping.
2. Understanding cartographic crowdsourcing In this section we seek to make sense of the phenomena of crowdsourcing for the production of cartographic information in the context of Web 2.0 developments by drawing upon the ideas of sociologist George Ritzer, and his notion of ‘prosumer societies’, geographer Michael Goodchild’s conception of volunteerist ‘citizen scientists’, and economic commentator Nicholas Carr’s critique of the ‘ignorance of crowds’. Taken together, we posit that these three theoretical positions make clear the attraction, power and problems of crowdsourcing with respect to the changing nature of the mapping experience.
2.1 Prosumption … the producers of much that exists on Web 2.0 are simultaneously the consumers of what is produced and there is a fluid relationship between production/consumption and producers/consumers. In terms of the latter, this is the emergence of the ‘prosumer’ (Ritzer 2008, 8). One way to begin to think about the practices of mapping and the changing culture of cartographic information production through crowdsourcing is to conceptualise the era of Web 2.0 as a distinct, new mode of capitalism. This is the thesis of U.S. sociologist George Ritzer in his theory of the ‘prosumer society’ (Ritzer 2008; see also Ritzer and Jurgenson 2010). Ritzer seeks to explain contemporary society by charting the evolving forms of capitalist economic relations. The dominant forms of social organisation are, he contends, the result of ongoing economic struggles between the forces of the market that seek profit and democratic ideals founded on notions of equity. The struggle is played out most overtly around labour power and the relations through which exchange value occurs. Ritzer (2008) makes a convincing case that with the rise of Web 2.0 a new mode of labour and exchange has started to emerge, what he terms ‘prosumption’. Prosumption is a form of consumption in which people fulfil a vital role or add crucial value in the creation or delivery of a service. In the main, prosumers do this additional work for little or no recompense, either getting enjoyment from the task, or a sense of empowerment, or they save money/time as the cost of the service is reduced and often becomes more flexible in nature (such as online banking or checking-in to a flight from home). In return the service provider or retailer receives free labour, talent, expertise, opinions, knowledge and gains efficiencies and potentially valuable information by getting ‘closer’ to the customers/clients (Ritzer 2008).
Prosumption capitalism is distinct in how it organizes and rewards labour power and restructures forms of exchanges and is thus different from production oriented (industrial) capitalism and the consumer capitalism that has been the dominant mode in the West for the past half century or so. The emergence of this new prosumer mode of capitalism, however, is not totalising and universal – it has not meant that production and consumption oriented modes of capitalism will be replaced and
disappear. Here we consider presumption and cartography with respect to (1) the selfservice economy, (2) the rise of amateur production, (3) and the prevalence of unfinished products.
It can be argued empirically that consumption practices have shifted quite significantly in recent decades as the self-service economy has broadened across many sectors. More significantly it has also deepened so that customers/clients must now routinely perform tasks for businesses/institutions that were previously the preserve of paid employees. This shift to self-service has reached a tipping point in the last few years with much greater automation offered through software (cf. Kitchin and Dodge 2011). People are increasingly becoming prosumers with respect to purchasing, payment, ticketing, billing and the tracking of goods. This is also the case with respect to mapping as online cartographic services enable people to become cartographic prosumers. Rather than rely on prepared cartographic products (like road atlases or paper street maps), prosumers create customised maps on-demand, able to set the frame, scale, content, legend, orientation, colour scheme depending on the platform. They can also access and layer multi-modal geospatial information, able to switch between and overlay map, satellite image, oblique views, street views and terrain displays, and actively query features, and in some case perform quite sophisticated comparative analysis (e.g., searching for new homes). A whole host of institutions and companies now exploit cartographic prosumption for efficient service delivery, compelling customers to answer their own spatial queries or to feedback and report incidents, such as road defects and vandalism to local government (e.g., ).
Many successful commercial firms seek to expand prosumption to be able to profit from distributing unfinished products that require the enrolment of the purchaser to complete them so that they function as intended. This has long been the model of ‘flat pack’ businesses such as Ikea’s self-assembly furniture. However, it has grown significantly in recent years, with many manufacturers no longer making standard products, instead shipping commodities that are bespoke, designed by customers via website options and assembled to order. Much software that is sold is the apeosis of an unfinished product, with many large applications offering multiple configurations, plug-ins, extensions and update patches, and are reliant on users to set-up their own 6
data structures and source key datasets. Installing base software from CDs does not get a working system. The mutability of computer code is in some senses one of the foundation stones of Web 2.0. Software – particularly that which is authored under opensource conditions – is an open-ended authorial tool. Its virtual form can be endlessly copied and easily configured so that it is almost infinitely adaptable. In the realm of cartographic information production, one might be an effective user of GIS software, but one gains new capacities by being able to upload and share GPS tracks or surveyed data to a web mapping platform such as OpenStreetMap, or to create a mapping mash-up, or being able to program and share code for an online mapping platform that enables new analytical and visualization functions. It is a shift from being a map reader to map maker.
Beside extending self-service interaction, another distinctive feature of prosumptive capitalism is the growth in ‘amateur’ labour power in various spheres, including media production and knowledge creation. The rise of the amateur is noteworthy both in terms of raw volume (more people volunteering in crowdsourcing) and also in terms of how this labour affects professional/paid workers by supplementing and replacing their contributions. This so-called ‘rise of the amateur’ is significant for cartography as it represents a fundamental change in the creation of maps. Rather than permitting people to interactively query and customize existing it enables them to prosume a new map into existence. Here there is a personalisation of maps, with the software and a locationally aware gadget able to beckon into being maps of the moment. This ego-centric cartographic production, most compelling on smartphones, allows amateurs to create their own ‘me-maps’ which are centred on their geographic location, updating in real time with their movement, and enabling the tagging of points, annotation and editing the cartographic information. Importantly, the code behind ‘me-map’ remembers its own mapping (in this sense it is a ‘logject’, see Kitchin and Dodge 2011).
2.2 Citizen scientists Michael Goodchild argues that one of the key features of neogeography – that is, new forms of geographical knowledge creation premised on Web 2.0 (see also Graham 2010) – is what he terms volunteered geographic information (VGI) (Goodchild 2007), effectively geo-crowdsourcing. For Goodchild, the spatial aspects of Web 2.0 7
– such as the range of new map mash-ups, free labour in authoring wiki maps, tagging of place-based information on virtual globes – is significant for future professional cartographic practice and map studies.
Goodchild (2007) observes that contemporary crowdsourcing can be seen as following longer established trends in ‘citizen science’ which are “communities or networks of citizens who act as observers in some domain of science” (p., 218). Here, people provide empirical observations and detailed measurements of phenomena for free, which are, crucially, of real value as data points for ‘proper’ science. Citizen scientists are particularly useful in observational terms because they can be geographically distributed and embedded in place through time (using professional research assistants to gather such field measurements is difficult to coordinate and costly). Examples include amateur meteorological observations made from back garden weather stations and amateur astronomers who persistently view and measure the night sky. Citizen science became well established through the twentieth century as a support to mainstream science, but remains a minority activity dependent on a fair degree of skill and knowledge or specialised equipment (such as a good quality telescope), and above all a considerable amount of motivation and commitment.
Taking a somewhat techno-centric position, Goodchild (2007) argues that for VGI to mature and extend further (with more people enrolled and performing more kinds of mapping) the barriers to entry must be lowered. This will be achieved, in part, by the routinization of simple digital tools for data collection, embedded in the functionality of mobile devices such as smartphones, and platforms for information authorship and collaborative knowledge creation becoming easier to use. Goodchild argues that VGI has value and should be nurtured because existing geographic knowledge about the world is incomplete and of variable accuracy. He points out that much of the world remains poorly mapped in terms of basic topographic survey knowledge, in spite of the availability of sophisticated GI tools and high resolution remotely sensed data. Where essential topographic mapping does exist it may not be wholly current – this is the situation in the USA and elsewhere. In some parts of the world that have been well mapped to date, it is not clear that this will remain the case into the future given the perilous resourcing of state-led survey infrastructures under neoliberalism. The danger is that cartographic information (based largely on data from state-funded 8
surveying) could conceivably regress with the decline in the status and capacity of state mapping agencies. And it is not a given that the private sector will step-up to fill in the gaps (Rhind 1997). Profit-driven mapping will not ensure universal, uniform coverage of space as companies will focus on profitable markets such as urban base maps that are relatively easy to capture, package and sell. Through VGI activities, it is anticipated, many more people can collectively act as a distributed set of spatial sensors, contributing, map data that could offer significant aggregate value. If only a small fraction of the six billion people on the planet is harnessed in generating new spatial information this could be hugely powerful set of citizen scientists.
Along with issues such as technological competence, user motivation and reward, VGI also raises issues of trust and democracy. If much more cartographic information is authored by the crowd, to what extent can we trust the process and the product? The trust and authority of traditional topographic maps stems from their author – the state. Nascent VGI projects, such as OpenStreetMap, clearly lack the capacity to demand authority and it remains to be seen whether people will trust their product. And while VGI promises to open up the authorship of mainstream representations of place, it remains unclear to what degree crowdsourcing is a form of democratic participation. Open-source mappers might have different characteristics to those employed by the state or commercial companies, but they are still a narrow and selective interest group and are not egalitarian or pluralist in organisation. The transition from closed to open-source cartography appears, on some registers at least, to be quite radical in formulation. However, existing power inequalities do not necessarily disappear, as the participation rates for VGI projects indicate, for example the analysis of Crutcher and Zook (2009) details how the ‘digital divide’ influenced the uneven spatial patterns in the generation of placemarks in Google Earth in postKatrina New Orleans.
2.3 Critique of crowdsourcing A third way of considering the amateur authorship at the heart of crowdsourced mapping is provided by economist and business commentator Nicholas Carr. He has set out critical argument focused on what he terms the ‘ignorance of crowds’ (2007). His work seeks to puncture the hype around technologies such as the Web, highlighting the potentially detrimental consequences flowing out to the real economy 9
from virtualised labour. He challenges the assumption that because sociotechnological assemblages like the Internet can provide new ways to organise and conduct existing activities that this will be inherently beneficial for society. He is critical, for example, of the way that Web resources and access to online media is active in changing how we think and create, arguing that the ‘internet is making us stupid’ (Carr 2010) in terms of how we engage with information.
In 2007 Carr critiqued the economic value in crowdsourcing, questioning the kinds of information that can be effectively produced by Web 2.0 approaches and the quality of mass amateur authorship. Accordingly, he argues that “peer production is best viewed as a means for refining the old rather than inventing the new; that it’s an optimisation model more than an invention model” (2007, 2). In other words, the value of Web 2.0 is not in producing raw content but in terms of being able to assemble an army of editors and checkers that can identify mistakes and improve large pre-existing information resources or projects. The potential of the crowd lies in correction and not creation. This kind of labour practice works because it can be easily subdivided and participants can undertake as little or as much as they want without detriment to the progress of the project or the overall quality of the end product.
These observations are interesting in terms of cartographic production where everyone is an expert on a locality, even if it is just their own street. If one enrols many eyeballs to check and refine existing maps, the labour could be quite potent. Commercial companies like Google and TomTom are beginning to go down the correction route by soliciting people to be prosumers and help spot errors and fix the map as they use the map. However, these companies seem less willing embrace a completely open model of amateur authorship; they need a degree of centralised command and control as they seek to sell mapping services for a profit.
Exploiting the crowd for purposes beyond checking and correcting according to Carr (2007) needs tight editorial control to assure quality of authorship. This is the fundamental tension in Web 2.0 between the attractions of open authorship and volunteerist labour and the continuing need for specific goals and guiding institutions.
For Carr crowdsourcing cannot be democratic or egalitarian as it risks quality, authority and usability. As he says:
“The open source model of production - when it works effectively - is not as egalitarian or democratic as it is often made out to be. Linux has been successful not just because so many people have been involved, but because the crowd’s work has been filtered through a central authority who holds supreme power as a synthesizer and decision maker. As the Linux project has grown, Linus Torvalds has gathered a hierarchy of talented software programmers around him to help manage the crowd and its contributions. It’s not a stretch to say that the Linux bureaucracy forms a cathedral that coordinates the work of the bazaar and molds it into a unified product.” (Carr 2007, 3)
He further illustrates this ‘ignorance of crowds’ through a trenchant critique of the ‘poster child’ of open, amateur authorship, Wikipedia. Despite its undoubted popularity1 and impressive amount of material authored in its relatively brief existence, Carr sees this particular knowledge product – and by implication the underlying approach – as deeply flawed. It suffers greatly from variable article quality which undermines the credibility and authority of the whole; there is no easy means of judging, from a superficial consultation, whether it is factually accurate. While some articles are clearly incomplete and poorly written, others may be flawed by much more insidious problems of biases and selective coverage. The result is an encyclopaedia that grows in haphazard directions with lengthy article on insignificant topics that happen to match the prurient interests, obsessions and populist fads of the small crowd of willing amateur authors. As Carr (2007, 4) notes “if Wikipedia weren’t free, it is unlikely its readers would be so forgiving of its failings.”
There are potentially deep consequences if Carr’s critical interpretation of contemporary political economy of authorship is correct when applied to the domain of cartographic information production. The willingness of voluntary mapping practices of amateurs will result not only in inferior VGI content, but also its free
In March 2011 it was ranked as the fifth most used website in world with 157 million visitors according Nielsen rating, .
status (both in terms of financial cost and also freedom of use) will negatively impact the profitability of cartographic companies and the viability of state agencies seeking to cover costs. A key concern is that for all its youthful energy and spirit of innovation, if OpenStreetMap really heralds the triumph of amateur labour in determining the exchange value for cartographic information that professional mapmaking will perish. It would also potentially devalue cartographic skills as people accept poor mapping products as the norm, and further erode the diversity of cartography. In addition, the mapping of certain things and places may become overlooked. In counter to such views, it is not at all clear that crowdsourced cartography is inferior to professional cartography (cf. Haklay 2010; Mooney et al 2011) or that opensource mapping will lack diversity, creativity and coverage. Indeed, it may well be the case that opensource and professional cartography might develop in new symbiotic ways.
3. The difference crowdsourced cartography makes Notwithstanding Carr’s critique of the value of amateur efforts, it is we believe worth considering the opportunities and benefits that might arise through crowdsourced and prosumptive cartography. Our focus of attention is OpenStreetMap2, probably the most well documented VGI project (Haklay 2010; Mooney et al 2011; Perkins and Dodge 2008; Ramm et al 2011; Zook et al 2010). More broadly we consider what cartographers and human geographers could gain by exploiting crowdsourcing in creative ways to do new types of research and perform different kinds of learning by doing.
At a practical and pragmatic level, spatial data produced by state and commercial companies can be expensive to access and its use bound by complex licensing deals, expensive recurrent subscriptions models and restrictive copyright clauses. For example, through the EDINA Digimap service, British academics can purchase institutional access to all current Ordnance Survey digital data for the UK, though it comes with strict licensing conditions concerning use and publication3. Such
The primary OpenStreetMap interface is at and further details on the project are given on the wiki, available at . It is also expensive – the University of Manchester pays £6,625 per annum for their Ordnance Survey Digimap subscription.
conditions limit academic freedom and constrict research. The ready availability of valuable but costly resources is likely to come under further pressure with widespread cuts in university and research budgets.
Beyond pragmatics of access and the financial sustainability, there is no doubt important research value in studying real world mapping practices as exemplified by crowdsourced projects such as OpenStreetMap. These projects reveal how processes of collaborative knowledge production occur in practice. As such, they enable a shift in the empirical focus of cartographic research from the rules and science of mapmaking and what maps represent and mean, to how maps are constructed in practice and do work in the world to solve different tasks. Such an epistemological shift also has the potential to open up larger intellectual questions about the nature of geographical knowledge in contemporary political economy with respect to authorship, the ontology of place, the partiality of representation, and temporality of production. These questions have the potential to connect academic cartographic research to other areas of contemporary social science scholarship such as material cultures, embodied practices, performativity of identity, digitisation of memories, surveillance and securitisation of everyday spaces and practices, mobility, and rhythm-analysis (cf. Dodge et al 2009).
3.1 Authorship In conventionally produced geographic knowledge, such as large, centrally organised cartographic surveys by the state, authorship is consciously rendered anonymous. Similarly, commercially produced street atlases and satnav route guides are also presented as authorless documents. Naming the authors would suggest subjectivity at play in map creation and denude the pretence of objectivity underpinning the correspondence between graphical signs and the territory represented. Here, there is an attempt to position cartographic representation as a single objective voice speaking onto space and not a polyvocal, subjective recantation that actually brings the space into being and which acknowledges the full range of multiple authors, with different skills and diverse subjectivities, that have contributed to processes of specification, data collection, checking, processing, drawing, labelling, deleting, designing, and so on. Such an acknowledgement exposes the map to suggestions of fallibility.
Typically opensource projects and ones that seek to enrol prosumer labour are more open about their authorship practices. In part they do this to attract the crowd and retain their labour by rewarding them through the positive acknowledgement and peer recognition of their contributions. In a wiki mapping project like OSM the authorship model by necessity fractures the anonymity of cartography products upheld by commercial organisations and state agencies. As of July 2011 some 441,651 thousand people had registered with OSM and 16,046 had contributed by editing – quite a sizeable crowd4. Although the average active set of prosumers labouring on the map during that month by uploading GPS data is much smaller at 3,954.
As with other VGI projects, in OSM all the cartographic information is created by registered users who become permanently associated in the database with the material they author and contribute. Most significantly, and unlike conventional cartographic production, this detail is publicly available, and can be queried and analysed. As a consequence, data on authorship can be analysed and mapped. Figure 1 presents an illustrative example, mapping the authorship of OSM for Bologna, Italy. The striking spectral presentation reveals the range of authors involved in producing this relatively small area. In total, the mapping crowd comprised 124 different registered users who have made an authorial contribution, although the top ten most active prosumers produced just over 74% of the map data being displayed. The dominant author is Alberto58, whose work constituted 18% of the data – coloured dark red, Figure 1 – which forms the dense core of the old city centre.
Figure 1. OpenStreetMap for the Bologna city area colour coded by the author of the data. The legend lists the authors in descending order based on the amount of data contributed. (Source: created by the authors in June 2011 using the ITO!’s OSM Mapper service, . A colour version is available from the Pion website.)
Whilst the primary web interface to the OSM database, rendered in the style of topographic cartographic artefacts appears similar in many respects to other commercial web mapping portals, crucially it has extra public functions (Figure 2). This includes an Edit button which is the gateway to the wiki based element of the project and allows registered users to change anything they choose. As the marketing
Source: self reported statistics for OpenStreetMap, .
slogan for OpenStreetMap states: “free, editable map … it is made by people like you”. Such open editing is a radical departure for cartographic products, shifting from top-down consumption to the opportunity for bottom-up prosumption.
Figure 2. The buttons the OpenStreetMap default interface that provides critical functionality for open authorship and also means the mapping processes can be studied. (Source: created by the authors in June 2011 from the website. A colour version is available from the Pion website.)
Selecting the View button and the data layer and then clicking on a data object in the map display, will query the record from the database and thus all details on the authorship. This ability to directly query any object and interrogate its properties is familiar to GIS users working on their own (private) database, but is powerfully new for a publicly accessible map of the world. The ability to see the complete technical history (through time and space) of contributors actions is clearly useful for analysing the processes of authorship (evidence of collaboration and conflict, how incremental improvements have emerged, the impacts of vandalism and the costs of their clean-up, etc). Figure 3 shows part of pattern of authorship for a single street in the centre of Bologna. Querying Via Marsala provides descriptive detail in terms of the feature tags associated with the object and more interestingly a full listing from creation to current status, showing every instance it has been changed with author name and date. It was first created in March 2008 and has been edited fourteen times over this period by seven different people (or at least registered users). It was last modified on 14th January 2010 by user Alberto58 (as of June 2011 when we last checked). Clearly, if one knows the area of the centre of Bologna it might be possible to comment on the completeness and quality of the map data in relation to reality and to focus in on the cartographic labour of specific prosumers to determine their reliability and therefore the authority of this instance of mapping.
Figure 3. Viewing a map object in OpenStreetMap, in the this case a street in the centre of Bologna, gives complete authorship history. (Source: created by authors in June 2011 from the website. A colour version is available from the Pion website.)
The rise of the amateur is a particular concern for cartography because it threatens to replace, not just supplement, professional surveyors and cartographers. Importantly, there is an element of deskilling going on – it takes much less effort and technical 15
knowledge to record spatial data with a recreational GPS unit than with sophisticated surveying tools. The use of amateur labour also raises issues of coverage, cost and currency of information production, as well as the integrity and authority of how space is recorded and what is deemed mappable. The danger with crowdsourced data is that it is subjective, emotive, biased and mired in the personal details, and what effective mapmaking really needs is the objectivity and synoptic viewpoint that can only be offered by a experienced, professional cartographers. What we have here are contrasting value judgements between the value of lots of local viewpoints versus a detached global singular overview. Walking a street, clutching a GPS may offer some visceral reality, but it does not necessarily unlock the truth of a place or generate an effective representation of space, which often only comes from setting immediate features into wider contexts. Arguably the real political significance of the rise of the amateur comes when the crowd can cut out the hegemonic (official) map itself to reach wider range of users directly. There is already little doubt that crowdsourced cartography can bypass existing power hierarchies and give opportunities for motivated prosumers to be in control of the message.
3.2 Ontology With a diverse range of mappers drawn as volunteers from the crowd, there clearly needs to be some sense of what of the world should be mapped in projects like OSM. This is a fundamental ontological question at the heart of geographic knowledge creation – what do we need to know about the world, what can feasibly be captured for a given purpose and represented in cartographic form? The new crowdsourced authorship of OSM exposes for public scrutiny such ontological questions and what people choose to capture and how they categorise and describe phenomena.
In conventional cartography, just as multiple authorship is denied, so too are the messy contingencies and selective interests embodied in ontological decisions. Such elisions work to bolster the notion of objective authority. As such, the politics of ontology relating to the selective interests of capitalist accumulation or militaristic agendas that underpin the ontological constitution of state produced maps, lies beyond the map itself. What, for example, is the ontological constitution of Google Maps? It clearly has one and it certainly arose from some debate within the company between GI experts, database specialists, interface designers and business managers, along 16
with the various contracted data suppliers, concerning what about the world should be mapped and how it should be presented.
The situation is often markedly different with open, crowdsourced, projects where prosumers are active in ontological discussions that mostly take place in online public forums. This is the case with OSM which has built an open ontology from scratch. Interestingly, many of the people most actively involved in the ontological development of OSM, whilst skilful and self-motivated, are not invested in conventional cartographic expertise and training, and are therefore not inculcated in existing (unwritten) ontological rules. As a result, there are often lively debates about the ontology for OSM (see Figure 4; these are also archived and can be analysed over several years5). With an open-source project potentially nothing is fixed and much is at stake. A good deal of reinventing occurs6 but also some fresh thinking from outsiders and non-experts as a new set of ‘rules’ for mapping the world through crowdsourcing and wiki-editing is coming into being.
Figure 4. An exemplar of ongoing ontological discussions for OSM that are conducted in public forums. Shown is a screenshot of part of the public archive of messages posted on Italian OSM list in January 2011. (Source: created by the authors in June 2011 from the material presented at .)
OpenStreetMap is a valuable ‘live laboratory’ to explore the ontological politics of cartography and the ways in which a mapping ontology appears solid, but is in fact always shifting. It is relatively easy to observe the multiple voices and the majority/minority positions in debates, and how contrary views are expressed and received by others, along with ways to deal with radically conflicting positions that cannot be reconciled with the existing dominant rule set. It is evident, for example, that OSM’s ontology is growing in complexity from immature beginnings to a fully fledged way of describing the world, but that this is causing tension amongst
These can be accessed at . Other freely available sources on the emergence of OSM’s ontology include the open editing of documents, FAQs and how-to guides on the wiki – especially the accepted feature list that sets up tagging conventions , regular open sessions on IRC and presentations by leading participants at the ‘State of the Map’ OSM conferences . OSM inevitably borrows much, often tacitly from existing cartographic norms, such as topographic conventions and a focus on roads.
contributors. Some want more things mapped – a significantly enlarged ontology – but others want a simpler ontology that might more easily be applied universally. There are also cross-cultural arguments about what counts as important to map and these are a particularly rich area to study. There have been a number of fascinating debates about naming conventions and what features in different parts of the world should be recorded and tagged (such as physical characteristics or its legal status). For example, the somewhat baroque nature of an English tagged ‘footpath’7 does not have the same meaning to most of the rest of the world. Here one can see how the Eurocentric core ontology, based on a developed physical national infrastructure, to consistent standards determined by the state, that dominated at the beginning of the OSM project is being challenged by contributors in other parts of the world, where these conventions make less sense.
Ontology is also tied to the politics of display and aesthetic conventions, where contributors tend to favour the national traditions they are most familiar with (Kent 2009 has shown the significance of national distinctiveness in basic topographic mapping through aesthetic judgements made by the state). There is a tension between those happy to have an ‘untidy’ ontology necessary to encompass local diversity, against those who support a universalist, all-encompassing ontology. Ontological discussions for OSM are also complicated because they overlap with issues of ownership over the data that volunteers are contributing and under what terms this should be licensed for others to re-use. This is particularly contentious in terms of commercial exploitation for profit of OSM data that has been volunteered for free. This is currently an ongoing debate and is unresolved.
3.3 Partiality of representations Crowdsourced projects like OSM are inevitably a work in progress and expose the partiality of all representations of place. Conventional maps, like a state-sponsored topographic sheet, or a street atlas on a satnav screen, assert the propositions of completeness and universality – a totalising god’s eye view of the world. Of course, they are not and can never be and this is a key deceit at the heart of the myth of cartographic objectivity. OSM exposes this deceit and allows it to be examined in
Including legal status and the particularities of bridleways versus byeways (cf. ) as well physical form.
detail. For example, the open API and free availability of the complete map database allow very different designs and presentation, and also support various comparative interfaces can present side-by-side or overlay different dates of OSM map data or juxtapose OSM data against other mappings. Figure 5 provides an example for Bologna, with OSM data and geographic imagery from another source in linked windows thus revealing the partial state of the OSM map or vice versa.
Figure 5. Comparative mapping interface demonstrating the partiality of OSM. (Source: created by authors in August 2011 using the ‘map compare’ tool provided by Geofabrik, .)
This ability to compare map data is of practical relevance in terms of creating better cartographic products, but it also explodes the persistent myth of maps as ‘mirrors’ of geographic space. The blank spaces on a map are not ‘empty’ but rather voids awaiting ascribed meaning (Harley 1992). Pedagogically, OSM’s incompleteness reveals how maps emerge through contested, negotiated relational practices, shaped in part by the socio-technical constraints of the measurement tools employed, such as those used to survey space. As critical cartographers have argued for more than two decades, cartographic information is derived not from the world in some pure and unmediated form, but is constrained by the parameters of capture technologies and filtered through the lens of what is deemed important by cartographers and their paymasters. In the case of VGI projects likes OSM, GPS enabled technologies play a key role in capturing position data. Whilst GPS appears to be a neutral technology that simply and accurately pinpoints location, it is rather a very particular way of knowing the world, one that comes bound-up with a military legacy (cf. Gregory 2010; Propen 2005), and has particular epistemic capabilities that empowers commerce and new forms of consumption (Caplan 2006).
Moreover, it has limits and its own partiality which are often ‘black-boxed’ from scrutiny. While it promises highly accurate measurements, the degree of precision varies across device, geographic context and user proficiency. As a result, the use of GPS in projects like OSM introduces degrees of imprecision and variability about location into the mapping database. Illustrative of this imprecision is the project by architect Laura Kurgan (1994) which exposed the myth of GPS precision by plotting
how an unmoving GPS unit records location points that vary around an average centre (Figure 6). As she notes this kind of work is important intellectually because:
“the space or the architecture of the information system that wants to locate and fix us in space has its own complexity, its own invisible relays and delays. The difficulty of charting the spaces that chart the spaces, of mapping the scaleless networks of the very system that promises to end our disorientation, demands redefining the points and lines and planes that build the map, and lingering in their strange spaces and times.” (Kurgan 1994, 17)
Figure 6. Illustration of the wandering nature of ‘precise’ GPS measurements made with a statically positioned unit. (Source: Kurgan 1994, 19.) Such critical thinking is relevant to contemporary work on digital geotechnologies and builds on earlier work questioning the embedded politics of GIS technologies (Curry 1998; Pickles 1995). As Brian Harley (1992, 231) argued at the start of the 1990s, digital cartography and GIS works “to strengthen its positivist assumptions and it has bred new arrogance in geography about its supposed value as a mode of access to reality”. GPS basks in an aura of precision, but for all its mundane sophistication it still produces a partial knowing of space. Working through the ‘nuts-n-bolts’ of generating cartographic information in OSM from raw GPS data reveals such partiality and imprecision. Consumer of maps typically do not dwell on how a map comes into being, but as prosumers become enrolled in their making, often from first principles, they become inevitably aware of the many subjective, relational and contingent decisions being made and the embodied and material practices of mapping with respect to cleaning, abstraction and generalization (see Figure 7). In conventional maps the ability to explore and question these practices is denied and the ability to examine raw measurement data is not permitted. But with OSM the full set of GPS ‘breadcrumb trails’ that people have uploaded and edited is available for consultation and quantitative analysis (cf. Mooney et al 2011) and interpretation.
Figure 7. Translating measurements into a map can be observed in OSM. Here editing of raw GPS tracks (a) and the proper geometry and tagging for cartographic information (b). (Source: created by authors from screenshots of the OSM website.)
With respect to the form of representation, Web 2.0 mapping, despite the possibilities presented by new technologies, largely reproduces existing cartographic conventions. For example, while OSM is a wiki that is open to many ways of mapping, the evidence is that recreates the same aspects of place that are mapped in state-sponsored maps and its semiotics are highly conventional. And yet there is much scope to push mapping in new directions. Some of the meta-maps deployed here (Figures 1 and 8), generated with the ‘hidden’ process data made uniquely open in OSM, point towards the possibilities. There is also scope for binding OSM with work by artists, for example, Barnes’ (2007) work on mapping spaces by recording graffiti or Nold’s (2009) work on charting the embodied emotional experience of space.
3.4 Temporalities of production The practices of mapping occur through time and yet conventional cartographic cultures tend to deny its temporality (e.g., a singular year date for print publication). As such, conventionally cartography representation has lacked temporal sensitivity to the unfolding of the world. Cartography produced by states, for example, portrays a static world in which dynamic, unpredictable and mobile processes are either ignored or frozen at specific and externally specified points in time. Beyond narrow technical concerns for legible representations, this freezing has important ideological implications because the resulting maps inherently privilege the status-quo and entrench power-relations that are most easily ‘frozen’. The surveyors gaze has always favoured things that appear fixed in the landscape (or with relatively long durations to human sensibilities). As Bunge (1971, 4) notes: “[m]aps attempt to integrate over time, that is, maps assume an average span of time. This means that nothing that moves is mapped, and therefore property is inherently preferred over humans.” Furthermore, most cartography conventionally displays selectively the material results of social and economic processes and not the processes themselves. For example, they map cities in terms of building footprints and not the activities occurring within those buildings or they map the spatial fixes of capital in detail (roads, factories, houses, fence lines, etc.), but not the underlying processes of unequal capital circulation.
OSM is interesting in that the underlying memory of how the map has become is kept in the database and can be interrogated. When one drills down to interrogate 21
individual spatial objects within OSM it is possible to see its complete genealogy in terms of authorial history and temporal emergence – when it first came into being, history of edits, and how current is its present instantiation is claiming to be (as illustrated in Figure 3 above). And, of course, one can then actually choose to edit the object to bring that part of the map into the present, and thus connect the viewing to the making of the space. This is the simple, but powerful notion of prosumption – the capacity of mapping in the moment of map use, which is certainly not possible with other cartographic products that are only consumable.
As with information authorship, so the patterns of temporal emergence of the OSM map can be rendered into their own analytical maps. As an illustration, Figure 8 displays the temporality for the mapping of Bologna. The majority of the city centre was fixed in middle of 2009 (coloured green) and has yet to be edited and updated further. We could imagine many other potential ways of seeing the memories of mapping practices encoded within the OSM database. Temporalities to explore might include: periodicity of large-scale change, durations, sequences, gaps in editing and bursts of activity. Many thematic analytical displays could be made of the oldest elements, which places are most frequently edited and updated, the spatial sequences and patterns of edits. And such temporal memories could be usefully cross-correlated to the nature of authorship in attempts to offer plausible explanations for varying prosumption patterns across time, space and time-spaces of the map. This might speak to the un-crowd like behaviour of crowdsourcing projects, such as sometimes relying on lone individuals and the lack of cooperation between contributors, as well as seeing how activity in OSM arises simply as an outcome of events on the ground (such as a news story about a place, a disaster, a proposal for major new development) and speaks to the issues of motivation and sustainability and the degree to which a community might come to care for a crowdsourced map of its meaningful places. Such temporal mapping would itself become a rich memorial device if it was to be continually changed over decades.
Figure 8. Temporality of OSM data for Bologna displayed as a thematic meta map. Such maps as memories explode the myth of temporal fixity that underlies conventional cartography. (Source: created by authors in June 2011 from the ITO! OSM Mapper service, www.itoworld.com>. A colour version is available from the Pion website.)
Conclusion In this paper we have examined the significance of Web 2.0 and crowdsourcing with specific respect to contemporary mapping experiences that underlay geographic knowledge production. Crowdsourcing is radically transforming the acts of making and using maps, fusing what were traditionally two separate practices, and altering fundamentally the political economy of the cartographic industry. We have argued that this is the case because Web 2.0 platforms and crowdsourcing participation engenders a new form of productive exchange, prosumption, and utilises the power of the crowd as citizen scientists. People are no longer compelled to purchase and consume maps. Rather they can access free to use mapping systems that enable them to prosume a map, able to customize the representational ‘look’ of the map interface, undertake elements of spatial analysis and, crucially, add to and edit the actual base cartographic information. They can act as ‘citizen scientists’, voluntarily contributing valuable spatial information to the mapping cloud, and then use this information to solve geographic tasks that matter. In so doing, we are beginning to see how prosumers can bypass the traditional processes by which maps are produced, published and consumed, and reconfigure the economic model of the cartographic data industry posing threats to state and private producers of data, especially given neoliberal rollback in state agency funding.
Moreover, the experience of crowdsourced mapping poses challenges to the professional authorship and normative ontology of cartographic representations and their supposed fixity, objectivity and authority. The production of crowdsourced mapping does not rely on professional training, experience and knowledge. Their message and ontology is not in the control of a single organisation or cartographer. Rather, they are explicitly polyvocal in authorship and the genealogy of author contributions, collaborations and conflicts are often publicly recorded and open to scrutiny. The message and ontology is collectively and openly negotiated and again this is subject to enquiry and available for analysis by scholars and others. In addition, crowdsourced ventures such as OSM reveal the partiality and politics of map representations and temporalities inherent in the mapping experience. . Such representational and temporal contingencies are typically suppressed within traditional cartographic production and distribution. Crowdsourced mappings thus 23
open up more fundamental ontological and epistemological questions about the nature of maps and mapping, a debate that has unfolded over the past twenty years with the development of critical cartography (cf. Crampton 2009; Elwood 2008). Despite the opportunities presented by crowdsourcing, some have questioned whether it can create information that is authored consistently, and with sufficient quality to match that produced by smaller numbers of paid professionals working in more conventional modes. Concerns relate to the evenness of coverage, consistency and accuracy of content and metadata, along with ongoing commitment and sustainability in labour exchanges without employments contracts and salary obligations. Quantitative analysis demonstrates that projects like OSM can produce cartographic information from field surveys that matches professional standards (Haklay 2010) and there is little evidence to date of ‘the ignorance of crowds’ feared by Carr (2007). Despite the rhetoric of mass involvement, it must acknowledged that OSM is crowdsourced by a few and not the many, with only a small active group who are setting the ontology and doing much of the mapping labour. In other words, there are a relatively small number of dedicated prosumers in comparison to the large number of consuming ‘free-riders’. Moreover, it could be legitimately argued that projects like OSM constitutes a huge amount of wasted energy in remaking what already exists simply to circumvent cost, copyright, and licensing restrictions.
Regardless of the various issues and critiques aimed at crowdsourced mapping, it is clear that cartographic data generation and map making as an industry has been fundamentally altered by opensource software technologies and Web 2.0 platforms. The mapping experience is being transformed. Whilst we have sought to explain the differences crowdsourced cartography makes, it is clear that more work needs to be undertaken to empirically research and think through and make sense of the profound changes taking place. Acknowledgements An earlier version of this paper was presented at the University of Bologna in June 2011 and we are grateful to Francesco Mazzucchelli and Patrizia Violi for the invitation to participate in their workshop. We also acknowledge the productive discussions with Chris Perkins, and the forbearance of the guest editors Matthew Wilson and Mark Graham.
References Barnes A, 2007, “Geo/graphic mapping”, Cultural Geographies 14: 139-47 Bunge W, 1971 Fitzgerald: Geography of a Revolution (Schenkman Publishing Company, Cambridge, MA) Caplan K, 2006, “Precision targets: GPS and the militarization of consumer identity”, American Quarterly 58(3): 693-713 Carr N G, 2007, “The ignorance of crowds”, Strategy + Business Magazine 47: 1-5 Carr N, 2010 The Shallows: What the Internet is Doing to our Brains (Atlantic Books, New York) Crampton J W, 2009, “Cartography: maps 2.0”, Progress in Human Geography 33(1): 91-100 Crutcher M, Zook M, 2009, “Placemarks and waterlines: Racialized cyberscapes in post-Katrina Google Earth”, Geoforum 40(4): 523-34 Curry M R, 1998 Digital Places: Living with Geographic Information Technologies (Routledge, London) Dodge M, Perkins C, Kitchin R, 2009, “Mapping modes, methods and moments: a manifesto for map studies”, in Dodge M, Kitchin R, Perkins C (eds), Rethinking Maps (Routledge, London), 220-43 Elwood S, 2008, “Volunteered geographic information: future research directions motivated by critical, participatory, and feminist GIS”, GeoJournal 72(3/4): 173-83 Goodchild M F, 2007, “Citizens as sensors: the world of volunteered geography”, GeoJournal 69: 211–21 Graham M, 2010, “Neogeography and the palimpsests of place”, Tijdschrift voor Economische en Sociale Geografie 101(4): 422-36 Gregory D, 2010, “Seeing Red: Baghdad and the event-ful city”, Political Geography 29(5): 266-79 Haklay M, 2010, “How good is volunteered geographical information? A comparative study of OpenStreetMap and Ordnance Survey datasets”, Environment and Planning B 37: 682-703 Harley J B, 1992, “Deconstructing the map”, in Barnes T J, Duncan J S (eds), Writing Worlds: Discourse, Text and Metaphor in the Representation of Landscape (Routledge, London), 231-247 Kent A, 2009, “Stylistic diversity in European state 1:50 000 topographic maps”, The Cartographic Journal 46(3): 179-213
Kitchin R, Dodge M, 2011 Code/Space: Software and Everyday Life (MIT Press, Cambridge, MA) Kurgan L, 1994, “You are here: information drift”, Assemblage 25: 14-43 Mooney P, Sun H, Corcoran P, Yan L, 2011, “Citizen generated spatial data and information: Risks and opportunities”, Proceedings of the 2nd International Conference on Network Engineering and Computer Science (ICNECS 2011), Xi'an, Shaanxi, China, 23-25 September, Nold C. 2009 Emotional Cartography: Technologies of the Self. O’Reilly T, 2005 What is Web 2.0: Design Patterns and Business Models for the Next Generation of Software, mimeo, Perkins C, Dodge M, 2008, “The potential of user-generated cartography: a case study of the OpenStreetMap project and Mapchester mapping party”, North West Geographer 8(1): 19-32 Pickles J, 1995 Ground Truth: The Social Implications of Geographic Information Systems (Guilford Press, New York) Propen A, 2005, “Critical GPS: Toward a new politics of location”, ACME: An International E-Journal of Critical Geographies, 4(1): 131–44 Ramm F, Topf J, Chilton S, 2011 OpenStreetMap: Using and Enhancing the Free Map of the World (UIT, Cambridge) Rhind D, 1997 Framework for the World (GeoInformation International, Cambridge) Ritzer G, 2008, “Production, consumption … prosumption?”, mimeo, Ritzer G, Jurgenson N 2010, “Production, consumption, prosumption: The nature of capitalism in the age of the digital ‘prosumer’”, Journal of Consumer Culture 10(1): 13-36 Zook M, Graham M, Shelton T, Gorman S, 2010, “Volunteered geographic information and crowdsourcing disaster relief: a case study of the Haitian earthquake”, World Medical and Health Policy 2(2): 7-33