Internet Geographer


Posts tagged knowledge
Uneven Geographies of User-Generated Information: Patterns of Increasing Informational Poverty (new paper)

After years of work, the first peer-reviewed paper to emerge from our research on Wikipedia is now officially ’in press’: 

Graham, M., Hogan, B., Straumann, R. K., and Medhat, A. 2014. Uneven Geographies of User-Generated Information: Patterns of Increasing Informational Poverty. Annals of the Association of American Geographers (forthcoming).

The paper has some very interesting and important findings, summarised in the abstract below:

Geographies of codified knowledge have always been characterized by stark core-periphery patterns: with some parts of the world at the center of global voice and representation, and many others invisible or unheard. However, many have pointed to the potential for radical change as digital divides are bridged and 2.5 billion people are now online.

With a focus on Wikipedia, which is one of the world’s most visible, most used, and most powerful repositories of user-generated content, we investigate whether we are now seeing fundamentally different patterns of knowledge production. Even though Wikipedia consists of a massive cloud of geographic information about millions of events and places around the globe put together by millions of hours of human labor, it remains that the encyclopedia is characterized by uneven and clustered geographies: there is simply not a lot of content about much of the world.   

The paper then moves to describe the factors that explain these patterns, showing that while just a few conditions can explain much of the variance in geographies of information some parts of the world remain well below their expected values. These findings indicate that better connectivity is only a necessary, but not a sufficient condition for the presence of volunteered geographic information about a place. We conclude by discussing the remaining social, economic, political, regulatory, and infrastructural barriers that continue to disadvantage many of the world’s informational peripheries. The paper ultimately shows that, despite many hopes that a democratization of connectivity will spur a concomitant democratization of information production, internet connectivity is not a panacea, and can only ever be one part of a broader strategy to deepen the informational layers of places.

This is the first of a handful of papers that are in the works, and I’ll post any updates that we have. In the meantime, feel free to get in touch if you have any comments, critiques, or questions about this contribution.
New paper: Geography and the future of big data, big data and the future of geography

Taylor Shelton and I recently convened an article forum on Geography and ‘Big Data’ in Dialogues in Human Geography.

Our lead article, “Geography and the future of big data, big data and the future of geography” (pre-pub version here) points to some of the potentials of 'big data’ in geography and some of the potentials of geography in 'big data.’ But it also argues that we need to remain cautions of the ways that 'big data’ can obscure, rather than reveal, the complexity of social and spatial processes.

We were privileged with a group of very knowledgeable and critical scholars to develop this important and timely conversation. The rest of the special issue can be accessed here, or through the links below.

Rob Kitchin: Big data and human geography: Opportunities, challenges and risks 

Evelyn Ruppert: Rethinking empirical social sciences 

Michael Batty: Big data, smart cities and city planning 

Michael Goodchild: The quality of big (geo)data

Sean Gorman: The danger of a big data episteme and the need to evolve geographic information systems 

Sandra González-Bailón: Big data and the fabric of human geography 

Trevor Barnes: Big data, little history 

'Data Shadows and Urban Augmented Realities' session line-up at the Annual Meeting of the Association of American Geographers

We have twelve papers confirmed for our Data Shadows and Urban Augmented Realities session at the next Annual Meeting of the Association of American Geographers. Full details below: 

Session Title: Data Shadows and Urban Augmented Realities

Session Organizers: 

Mark Graham 
University of Oxford
Oxford Internet Institute

Matthew Zook
University of Kentucky
Department of Geography

Session Abstract: 

Most parts of our urban areas have become both digitally connected and represented by digitized information. Digital layers of geographic information (commonly referred to as “augmented reality” by computer scientists) can take myriad forms. The most visible of which are probably the digital maps that many people use to navigate through cities. Google, Yahoo!, Wikipedia, Apple, OpenStreetMap, Baidu, and many other companies and organisations all host publicly accessible platforms that partially reflect parts of our world. These services also become the platform for an almost unimaginable amount of additional content that both reflects the materiality of cities and augments it with additional content. This additional volunteered (and emitted) geographic information is comprised of photographs, blogs, tweets, social media checkins, webcams, videos, and encyclopedia articles. These layers of digital representations are then further reproduced and repurposed in the ways that they annotate the urban environment


The ambition of this session is to interrogate the increasing prevalence of both geographically referenced digital information and the code through which it is regulated. By asking what these augmented realities are, where they are and where they are not, and how they are brought into being, we can both unpack the language we use to speak about digital augmentations and explore the ways in which digital extensions of place are becoming increasingly important in everyday, lived geographies.


This session seeks two kinds of papers. First it aims to provide space for papers that explore the ways in which we should imagine, describe, critique, and even name, the digital and informational augmentations of our lives. Second, the session seeks papers that critically examine information geographies and augmented realities in specific contexts. How do informational augmentations impact on how we bring our worlds into being? What and where do they exclude? What narratives and discourses do they allow, and what do they conceal? How are they governed, regulated, and challenged?

Session 1: Practicing Data Shadows


Visual data shadows and the difference between seeing and photographing designed landscapes


Erin Despard 

University of Glasgow

School of Geographical and Earth Sciences

In this paper I explore the significance of gaps in the augmented ‘reality’ created by the proliferation of images online and in social media, and the extent to which visual data shadows may serve as a means of mapping or interrogating cultural values. I am particularly interested in images of designed landscape, which tend to depict urban environments in a manner that is apparently diverse but highly selective. My case is the newly renovated Grand Park in Los Angeles, where a major feature of the design and programming is the inclusion of gardens narrating a natural history of trees in LA. While these gardens present an opportunity to see the city differently—that is, as an urban forest populated by species other than palm trees, planted by a variety of cultural groups—they are relatively absent from photographs circulating via Flickr and Instagram. There are many ways of accounting for this invisibility—including the design of the landscape itself, which prioritizes views unobstructed by trees—but I focus on finding a means of making this gap in the site’s visual augmentation appear, while also theorizing its significance. What is the relationship between the circulation of images and ways of perceiving the park, and what does that in turn suggest about how visual augmentation in general works? A lack of photographs suggests that people are not attending to a given feature, but we need to think in terms of production practices and the differing contexts of circulation in order to fully unpack the significance of that inattention. 


June 16th 1904 in Dublin: Augmented Realities and James Joyce’s Ulysses(1922)


Charles Travis

University of Dublin Trinity College

Trinity Long Room Hub



The digital layers of augmented urban geographical information hosted in Google, Yahoo!, Wikipedia, Apple, OpenStreetMap, Baidu, are ubiquitously engaged to navigate the world’s cities and have diffused into the sphere of literature.  Michael Connolly and Lee Child both employ these platforms in their crime thrillers and in Joseph O’Neill’s Netherland (2008) Google Earth is used by its Dutch narrator as a navigation device to vicariously remedy home sickness while living in the Chelsea Hotel in Manhattan. Indeed, the relations between accessible geographical information and the construction of fictional narratives is a long standing one. A century ago urban information was commonly compiled in city gazetteers supplemented by foldout maps.  One example is the 1904 edition of Thom’s Official Directory of the United Kingdom of Great Britain and Ireland, employed by James Joyce while writing Ulysses (1922) to “remotely sense” from Trieste, Zurich and Paris the layout of Dublin’s streets, districts, pubs, churches, houses, and neighbourhoods, as he plotted the journeys of Leopold Bloom, Stephen Dadelus and other characters as they crossed the city on the 16thof June.  Frank Budgen recalled that “to see Joyce at work on the Wandering Rocks” section of the novel “was to see an engineer at work with compass and slide-rule, a surveyor with theodolite and measuring chain, or more Ulyssean perhaps, a ship’s officer taking the sun, reading the log and calculating current drift and leeway” (1972, p. 123). Joyce’s literary navigation of Dublin was also informed by the visual techniques of Cubism, Futurism, and Dadaism. Budgen points out that “the multiplicity of technical devices is proof that Joyce subscribed to no limiting aesthetic creed and proof also that he was willing to use any available instrument that might serve his purpose”(1972, p. 198). One can only imagine if Joyce had access to the augmented realities produced today by GIS / GPS / Google Earth / Open Source platforms to navigate and explore Dublin as he creatively repurposed the city’s digital data to plot Ulysses.  This paper has two aims. Firstly, it will try to counter-factually envision a scenario which finds Joyce Googling information, accessing Floating Sheep and playing with Thom’s map of Dublin in ArcGIS, as a means for us to interrogate the process of creating further fictions by gleaning information from the digital augmentation of Dublin’s urbanity. Secondly, it will juxtapose this counterfactual with a discussion of the Bloomsday app JoyceWays, and other digital platforms related to Joyce’s work, to explore within the specific context of these augmented realities which Joycean narratives and discourses they allow, which ones they conceal, and how their digital platforms frame the imagined geographical and aesthetic journeys through the periodicity and pages of Ulysses, as a reader simultaneously navigates the streets of contemporary Dublin.


UPLOAD: Urban Politics of London Youngsters Analyzed Digitally


Koen Leurs

London School of Economics and Political Science (LSE)

Department of Media and Communications



The main aim of the proposed study is to investigate the lived experience of cultural difference among young Londoners (between 12-18 years) of different cultural backgrounds. Internet applications such as the video sharing platform YouTube, the social-networking site Facebook and micro-blog Twitter are taken as entry points to study the juxtaposition of differences in urban, digital representations. I will theorize and produce new empirical knowledge about how digital practices become loci of intercultural encounters. Taking a comparative approach, I focus on the networked belonging of youths from lower-class (often more multicultural) and upper-class (often more homogeneous) London boroughs on FacebookTwitter and YouTube. As digital practices have become a significant part of their life, it is urgent to achieve greater insights in whether their use of Internet applications corroborates pan-European sentiments of failed multiculturalism and ethnic segregation or whether their experiences rather showcase conviviality, cross-cultural exchange and cultural hybridization. Thus far, the ways in which diverse ethnic/gender/religious identities digitally encounter, negotiate and appropriate one another across online/offline spaces have remained understudied. Innovatively bringing new media, gender and postcolonial studies into dialogue; the layered dynamics and user-generated cultural heterogeneity across Internet applications is scrutinized. The proposed study combines large-scale digital methods to study geographically tagged user-generated content, qualitative in-depth interviews with 90 youths and virtual ethnography with 30 young informants.


Blanked Out: Data-rich and Data-blank Spaces in London as a New Frontier of Gentrification



Lilas Duvernois-Guevara and Gili Vidan

University of Oxford

Oxford Internet Institute



This paper investigates the sources of disparity in digital data collection on user-generated mapping platforms in the City of London. Reexamining the phenomenon of gentrification in a digital context, the study measures participation in the augmented urban environment by analysing  user-generated geographic information. Critical geography literature illustrates how inequalities become embedded in the architecture of cities. Yet, as urban environments are increasingly represented digitally, there is a growing need to account for the conditions that enable disparity in generating this urban-geographic data. Collecting metadata for businesses and points of interest, using Google Maps, Yelp and Foursquare, this paper correlates those distributions of geographic data with economic census indicators to produce overlays of digital presence and urban social processes. The reality represented by this location-enabled data is then compared with the experienced space through field surveys of three neighborhoods. This ground truthing process includes physical surveys of businesses and qualitative interviews with residents.



We propose a new analytic dimension to the urban phenomenon of gentrification—by presenting lack of inputs as blank spaces, user-generated mapping amplifies social divides in the augmented urban environment. We conclude that augmented realities are sensitive to a highly particular set of data-generators. The sensors measuring activities in these augmented realities are tuned to specific inputs, commonly aggregated by a particular socio-economic group. Through this process, those coded spaces generate a sub-group of urban residents digitally unaccounted for. In addition, this paper contributes to the research of digital spaces by devising a categorization of spatial data’s richness and blankness.

Session 2: Coding Data Shadows


Augmented cities, socially mediated geo-platform and the geo/coding of places 

Sung-Yueh PerngTracey P. Lauriault and Rob Kitchin 
National University of Ireland Maynooth
National Institute for Regional and Spatial Analysis


This paper seeks to understand augmented cities through the interplay of processes of abstraction and geo/coded specificities.Processes of abstraction in computing are part of architecture building and code writing. It is the practice of formalizing the relationship between various entities observed and the directing of how these behave once an algorithm is set in motion.  In this instance, once a social media geo-platform is in play, mobile users geo/code places within the location-based framework become determined by the program. For example, in Foursquare geo/coding occurs as individuals 'check-in’ by locating themselves in the programmed version of the ‘city’,  and then geo/code it with updates, comments and multimedia. This is therefore the remediation geo-referenced information, that was exchanged by word of mouth and paper maps, to real-time positioning and mobile networking. Once these elements are programmed into the ‘city’, it becomes an engineered promise of enmeshed and diversely experienced 'locals’.  The knowledge of the city for better or for worse is augmented by this practice.


The assumptions built into geo-platforms, is that they can be represent anywhere and everywhere. This presumes that cities, the specific entities within them and the relations among these are interchangeable.  Concurrently, geo-platforms are continuously being updated, functionality is added and often external actors build new apps and create services to extend them.  This questions the initial assumption that one geo-platform fits all places and that places have no specificities.  Meantime, there is the belief that a city’s specificities can be encapsulated in a tweak and/or a hack.  The recognition of this tension around processes of abstraction and specificities, within and outside location-based frameworks and the geo/coding of places these allow, highlights the misplaced promise and the hope for the seamless alignment between programmed dreams, living in a and the augmented city. 



The semantic production of space: pervasive computing and the urban landscape


Matthew Kelley

University of Washington

Urban Studies and Geographic Information Systems

Abstract: This paper suggests that as pervasive computing technologies have gained purchase in urban space they have also become more implicitly blended with everyday life and more contingent on information that is inductively compiled from Internet-based data services. It is argued that existing theorizations of the technologically mediated production of urban must engage with the increasingly implicit nature of informational transactions as well as the emergent semantic structuring of information. Drawing on examples of ongoing pervasive computing projects, implicit computing procedures are explored in relation to the mediation of everyday urban life. Literatures from computing science and geographical theory are brought into conversation in order to examine the consequences of a convergence between implicit pervasive technologies and the spaces of everyday life.


From Jerusalem to Kansas City: New geopolitics and the Semantic Web


Heather Ford and Mark Graham

University of Oxford

Oxford Internet Institute



The world’s most popular search engine, Google, has recently announced a fundamentally new way of representing populated place. With its ‘Knowledge Graph’, which is a ‘semantic network’ containing over 18 billion facts, the search engine is changing how geographic information is presented. Conducting a search for “capital city of Israel”, for example, currently results in the prominent display of facts about Jerusalem including a map with its borders, size, population numbers, weather and points of interest.


Google’s embrace of the Semantic Web was seen as a major success for the movement calling for the next moment in the Web’s history - a moment in which information is being increasingly tagged, standardised, and ordered in order to allow information to be more easily shared and reused. But this fairly innocuous-sounding movement has a single impact that many didn’t foresee. Categorization of cities is taking place outside of the control of those affected by those sites and is now in the hands of a technical elite.

This paper tracks data about two cities as they appear, congeal, morph and are transfigured by its movement through three of the world’s most important data platforms: Wikidata, Wikipedia and the Google Knowledge Graph. In doing so we ask three key questions: In whose interests are these snippets of data being decided? What are the material conditions that have led us to this point in the history of the web? And what does it mean for the future of cities and the voices that represent them? 


Code and the mediation of Space: The pleasures of being placed


Jim Thatcher

Clark University

Department of Geography



As public space has been overlaid with invisible, electronic networks made up of wireless media and devices, myriad aspects of society have become enmeshed with and mediated through devices capable of computing – of calculating and processing data. The very ubiquity of code as a mediator in human’s lives has led to its disappearance from conscious consideration. This “invisible infrastructure” of hybrid spaces reshapes the experience and function of late capitalist modernity. While technology has long been associated with radical changes in culture, politics, and economics and the specific mobilities and dynamism offered by information networks has been widely acknowledged, less has been said on how mobile, digital, spatial technologies are shifting the social construction of space. While space remains socially constructed as Doreen Massey suggested, this research sheds light on whose social experience is encoded into software and comes to mediate lived experience. Drawing on qualitative interviews and observations of individual users of mobile, spatially-aware applications – such as YelpFourSquare, and others – this talk highlights the role code plays on the social construction and experience of space. Being placed on a map, and in a system of consumption, engenders feelings of confidence tailored to the desires of application designers and the needs of large corporations. While retaining ultimate agency over their location, end-users enmesh their digital ‘location’ within a system of data’s creation, commodification, and exchange.

Session 3: Tracking Data Shadows 


The Digital Divide in Volunteered Geographic Information


David A. Parr

Texas State University-San Marcos

Department of Geography



As the Internet has matured, new technologies have blurred the line between data creator and consumer while decentralizing the creation of content. The term Volunteered Geographic Information was coined to describe the contribution of online geographic data by citizens (Goodchild 2007). The rapid increase of available online data including VGI has been labeled the exaflood. Even while the amount of data online is increasing, a 2013 Pew Research Center study found that 15% of U.S. adults do not use the Internet (Zickuhr 2013). The production of VGI is believed to fall along the gradient of skills, access, and technologies that form the digital divide We further believe that there is a geographic pattern to this digital divide on VGI and that the pattern can be mapped and predicted by the distribution of key variables. This paper will report on our examination of VGI/digital divide from two aspects - (1) how is the production of VGI related to data contributors, and (2) how does the digital divide of VGI illustrate itself spatially in the US? The objectives of this study are to evaluate a large VGI dataset using 'big data’ methods with the aim of understanding the uneven spatial distribution of VGI data. Using data extracted from OpenStreetMap, the paper will compare and map the top fifty 2010 Metropolitan Statistical Areas for OpenStreetMap activity from 2005-2013; it will relate county-level socioeconomic digital divide variables with OpenStreetMap activity to evaluate if there is a significant relationship between income, education, and OpenStreetMap activity.



Measuring Digital Inequality Through Social Media Data Shadows?


Claudio CalvinoGrant Blank, Stefano De Sabbata and Mark Graham  

University of Oxford

Oxford Internet Institute

The Internet is an increasingly important part of British social and economic life. Britain has the largest Internet economy in the industrial world, measured as a percent of GDP. Social life is increasingly mediated, influenced, and augmented by online interactions that take place through the Internet. Yet despite the importance of the Internet in everyday life, we know surprisingly little about the geography of Internet use and participation at sub-national scales.

The 2013 Oxford Internet Survey (OxIS), a random probability sample of 2,657 households in Great Britain, asks many questions about digital participation, interests, activities, and the types of content that people produce. However, the sample size of the survey is too small to make detailed inferences about the geographies of digital inequality below the regional level. By combining 2013 OxIS data with social and economic statistics from the 2011 census, and social media data from a database of Twitter, we can use small area estimation techniques to estimate Internet participation and other variables down to the county level for all of Great Britain. This allows us to map the geography of internet use at levels heretofore unknown. More theoretically, ask how we might expand some of the insights from OxIS using traditional (the census) and unconventional (social media data shadows of places) means. Ultimately this work leads us to ask what methods are appropriate to measure local-scale geographies of digital participation. 



An analysis of the resilience of augmented realities to news events

Stefano De Sabbata and Claudio Calvino

University of Oxford

Oxford Internet Institute



This study investigates the intersection of various kinds of media in the contexts of augmented realities at moments of stress. We specifically ask whether geosocial media reflect the same patterns that are produced through more traditional producers of media (i.e. newspapers). In doing so, we can better understand the diverse layers of content that augment our experiences in moments of conflict, stress, and unrest.  To do this, we consider the response of the Web to a 2013 terrorist attack at the Westgate shopping mall in Nairobi. We will conduct a quantitative analysis of two datasets: first, the “Global Database of Events, Language, and Tone” (GDELT), which is a collection of 250 million stories published in newspapers from almost every country on Earth; and second, a collection of geocoded tweets extracted from a set of 2 million collected over the course of a month.




In both databases, we identify locations and events which mention the attack, and analyse the geographic and temporal distribution of items in both datasets. Specifically, we compare engineering resilience (recovery times), raster-based distributions, and aggregations to larger spatial units. In doing so, seek to understand whether global and local attention (over time and over space) in the data shadows of this event differ substantially. Drawing on the literature on resilience and spatial perceptions, we then attempt to ask what the effects-in-the-world of these variable geographies and temporalities might be. 



To Tweet or not to Tweet? Exploring the Demographic Data Shadows of Twitter


Matthew Zook and Ate Poorthuis

University of Kentucky

Department of Geography


Geosocial media provides new ways to understanding places, particularly how they change over the course of a day as well as documenting connections between places.  These insights, however, come with heightened privacy worries and concerns about the biases contained within the data shadow associated with each particular form of social media.  In short, we know far too little about the representativeness of geosocial media data and are just beginning to grasp the implications on privacy associated with its use.

This paper seeks to fill this gap with an intensive study of one form of social media data, geotagged Tweets.  Building on the DOLLY database, a repository at the University of Kentucky containing all geocoded tweets since December 2011, this paper quantifies the demographics of Twitter users.  Using a dataset based on a random sample of 100,000 American Twitter users and the ~55 million tweets they sent in the last 16 months, we explore a number of techniques to establish home location of users, such as sustained presence and movement. Once home locations are identified we draw upon demographic data from the U.S. Census at the block group level to provide some insight on the nature of these users. While ever mindful of ecological fallacy, we explore the relationship between the density of inferred home locations and the demographics of that geography. This also allows us to go beyond a demography based on where people sleep and instead combine these demographic profiles with the intra-day movements of each user. In doing so, we document connections between locations and demonstrate how the demographic profile of any particular location changes over the course of a day.


The goals of this paper are 1) to outline a method for connecting geosocial media with official demographic data and 2) use geosocial media to better understand the demographic dynamism of locations in the course of day.

geography of top-level domains

We’ve just published a new map over at the Information Geographies project. This one draws on some of Matt Zook’s research into the geography of domain names. The map offers a detailed overview of one-facet of the geography of content production.

Some results are unsurprising (for instance the low scores in many countries that have low numbers of internet users). However, other unexpected patterns also reveal themselves (such as the relatively low numbers of domains in many Asian countries).

For a more detailed description of results take a look at the discussion on the map’s Internet Geography home.