Wednesday, July 22, 2015

New job working with the Geonet team at the Oxford Internet Institute: 'Researcher in ICTs, Geography and Development'

We are now hiring a researcher to work with us to investigate low-wage digital work being carried out in Sub-Saharan Africa.

The Oxford Internet Institute is a leading centre for research into individual, collective and institutional behaviour on the Internet. We are looking for a full-time Researcher to work with Professor Mark Graham on the ERC-funded project Geonet: Investigating the Changing Connectivities and Potentials of Sub-Saharan Africa's Knowledge Economy. Combining archival research, surveys, and in-depth interviews, this ambitious project will critically assess the changing landscape of digital work in Sub-Saharan Africa, and ask who benefits (and who doesn’t) from those changes.

In this exciting role, the Researcher will carry out 9-12 months of fieldwork among digital workers and organizations in Sub-Saharan Africa, as well as working at OII’s premises in Oxford. The Researcher will also contribute to the dissemination of the findings through peer-reviewed academic papers, project reports, events, blogs and social media.

Candidates should have experience of social science research in Development Studies, Geography, Sociology, Social Anthropology, Communications, Organization Studies, Management or related disciplines, training and practical experience in qualitative research methods.

Based primarily at the Oxford Internet Institute (with periods of fieldwork), this position is available immediately for 3 years in the first instance, with the possibility of renewal thereafter, funding permitting. For qualified candidates, there may also be opportunities to teach course modules on our ‘Social Science of the Internet’ MSc course.

The application form and further details, including a job description and selection criteria, are available on Oxford University's recruitment website.

The closing date for applications is 12:00 BST on Thursday 3 September 2015 and only applications received before then can be considered. Interviews for those short-listed are currently planned to take place in the week commencing Monday 21 September 2015.

Tuesday, June 9, 2015

Uneven Geographies of Digital Wages

Our previous post contained a few maps that shed light on the geographies of online work. But what we didn't do there was explore the spatial variance of wages.

The cartogram in this post depicts each country as a circle and sizes each country according to the wages for digital work flowing into each country every month (on the platform). The shading of the inner circle indicates the median wage for digital work in that country. The graphic broadly reveals that median wages are, perhaps unsurprisingly low in developing countries and are significantly higher in wealthier countries.

With my colleagues Isis Hjorth, Vili Lehdonvirta, and Helena Barnard, working on a multi-year project on digital labour and development, we'll have have some research coming out soon that explores some of the reasons why digital workers in the Global South have less bargaining power than their counterparts in the North, and why workers in the North can often attract a premium for their work. The issue is not just a global race to the bottom (which in some ways seems to be occurring), but that there are different price floors for wages in different parts of the world, and there are wage penalties associated with being affiliated with some countries.

(Elance-oDesk Inc. supported this work with their data; our project is funded by the IDRC; also thanks to Stefano De Sabbata, Claudio Calvino, and Sanna Ojanpera for their help with this work. This is a cross-post from the Connectivity, Inclusion, and Inequality website)

Monday, May 11, 2015

Code, Content, and Control: Global Geographies of Digital Participation and Representation

The Open University has just posted a webcast presentation of our paper due to be delivered at ICCG 2015 this July.

Titled “Code, Content, and Control: Global Geographies of Digital Participation and Representation”, it deals with the role of Google in the reproduction of place-based inequalities. It does this thought an examination of how digital information mediators command power through the abstract representation of space. It then ends by asking what alternatives look like; and how we actually might go about getting rid of monopolists of spatial information (i.e. Google). 

As part the Open University's Annual Doreen Massey event on digital geography, our talk provided the concept note for the section on "reformulating development geography" which also featured interventions from Sarah Elwood and Valentina Carraro from Grassroots Jerusalem. The entire event is available here and provides a great overview of some of the key contemporary debates in digital geography.

The Domestic Turn: Business Processing Outsourcing and the Growing Automation of Kenyan Organisations

I'm happy to announce a new paper to come out of our previous project studying Development and Broadband Internet Access in East Africa. The project was a collaboration between myself, Tim Waema, Laura Mann, and Chris Foster and aimed to look at the role that changing connectivity in East Africa was having on three sectors of the economy: tea, tourism, and business process outsourcing (BPO).

This paper, written with Laura Mann, focuses entirely on the BPO sector in Nairobi. There were high hopes that better connectivity would allow Kenyan firms to attract BPO business from around the world. However, the paper demonstrates that something significantly different has happened. Firms increasingly see Kenya as a market rather than as a base from which to build a globally competitive BPO sector. This work will thus inform much of what we do as we begin fieldwork in the GEONET project.

A pre-print of the full paper and abstract can be accessed below:

Mann, L., and Graham, M. (2015). The Domestic Turn: Business Processing Outsourcing and the Growing Automation of Kenyan Organisations. Journal of Development Studies. (forthcoming).


After observing the growth of the Indian and Filipino Business Processing Outsourcing sectors, Kenyan policy-makers and managers made substantial investments in international internet infrastructure and BPO marketing campaigns. While observers continue to discuss the sector in terms of its international work opportunities, in recent years the sector has increasingly focused on contracts sourced from Kenyan and other East African clients. The government has also refocused efforts on attracting international BPO companies. This domestic turn signals both the difficulties of gaining access to overseas work due to the power of incumbents and the increasing use of the internet and ICT-enabled automation within Kenyan organizations. In effect, better connectivity has enabled a two-way globalisation of services: Kenyan BPO companies have been able to access some international work opportunities but the connectivity has also contributed to the inflow of international service companies and business practices into Kenya. The conclusion examines what these shifts might entail for the sector and its workers in future.

Wednesday, April 29, 2015

The hidden biases of Geodata

The Missing Ground-truth in Geographic Data - Figure 1
Geographic information underpins so much of what we do today on the internet. By knowing the location of a tweet, a profile, or any other user-entered information, we can build services and software that is micro-targeted at user needs: for example dating sites, advertising, and search results.
For that reason, Stefano De Sabbata and I have done some work to understand the biases embedded in the databases that we use to create geographic ground truths. 

I've written up some of the findings for a post in The Guardian (Graham, M. 2015. The Hidden Biases of Geodata. The Guardian Apr 28, 2015.), and you can read more about the research in a pre-print of our forthcoming paper:

Graham, M. and De Sabbata, S. 2016. Mapping Information Wealth and Poverty: The Geography of Gazetteers. Environment and Planning A. (in press).
The Missing Ground-truth in Geographic Data - Figure 2

Thursday, February 12, 2015

New paper: "Barriers to the Localness of Volunteered Geographic Information"

Screen Shot 2015-02-12 at 09.48.45 copy

Some colleagues (Shilad SenHeather FordDave MusicantOliver KeyesBrent Hecht) and I have put together a paper for CHI on Barriers to the Localness of Volunteered Geographic Information. The paper asks important questions about both the geographies of information, and the factors that explain those geographies: 
Localness is an oft-cited benefit of volunteered geographic information (VGI). This study examines whether localness is a constant, universally shared benefit of VGI, or one that varies depending on the context in which it is produced. Focusing on articles about geographic entities (e.g. cities, points of interest) in 79 language editions of Wikipedia, we examine the localness of both the editors working on articles and the sources of the information they cite. We find extensive geographic inequalities in localness, with the degree of localness varying with the socioeconomic status of the local population and the health of the local media. We also point out the key role of language, showing that information in languages not native to a place tends to be produced and sourced by non-locals. We discuss the implications of this work for our understanding of the nature of VGI and highlight a generalizable technical contribution: an algorithm that determines the home country of the original publisher of online content.
You can access a copy of the paper here:

Sen, S. W., Ford, H., Musicant, D. R., Graham, M., Keyes, O. S. B., Hecht, B. 2015 Barriers to the Localness of Volunteered Geographic Information. CHI 2015 (pre-publication version here).

And you can also watch a screencast (excellently narrated by our own Heather Ford) of our forthcoming interactive map tool:

Wednesday, January 21, 2015

Informational Magnetism on Wikipedia: mapping edit focus

The previous post demonstrated not only that Sub-Saharan Africa and the Middle East and North Africa are net-importers of content on Wikipedia (Sub-Saharan Africa, for instance, receives 10.7 more edits from the rest of the world than it commits to the rest of the world), but it also showed where those edits come from.

This post does something a little different: it shows where edits are sent to.

Screen Shot 2015-01-21 at 09.37.32
In this network, an edge (a line connecting nodes) thickness is proportionate to the total number of edits received. That is, if a region sends most of its edits to North America (even if it sends very few edits) then that edge will be thick.

With the graph normalized by the number of edits sent, we see a striking finding from the Middle East and North Africa (MENA) region. Not only do less than half of edits in the MENA region stay within the MENA region, but a substantial share go to North America, even if their impact on the overall number of edits within North America is rather minimal.

In fact, proportionate to the total number of edits sent from a region, MENA sends more edits to North America than any other region sends to another. This leads to what might be considered a ‘double whammy’ for the MENA region. Not only are there not many articles about the Middle East, and even fewer in local languages. And not only are there very very few editors. But of the edits that exist, a lot of them are to write about Europe, Asia, and America.

The global informational cores are exerting a sort of information magnetism. The presence of information creating a virtuous cycle of informational richness; and the absence of information being part of a vicious cycle of informational poverty.

Ralph Straumann and I are working on a paper that explores these topics and this conclusion in more detail, and will post a full draft here as soon as we have one. In the meantime, we’d welcome any comments or questions on the patterns and data presented here.

Note that this work comes from the following report:

Graham, M., and Hogan, B. 2014. Uneven Openness: Barriers to MENA Representation on Wikipedia. Oxford Internet Institute Report, Oxford UK.

The following paper also offers an abridged version of some of the results:

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).

Or for a broader discussion about why the locality of participation matters, see:

Graham, M. 2014. Internet Geographies: Data Shadows and Digital Divisions of Labour. In Society and the Internet: How Networks of Information and Communication are Changing our Lives, eds M. Graham and W. H. Dutton. Oxford: Oxford University Press. 99-116.

Graham, M., M. Zook, and A. Boulton. 2013. Augmented Reality in the Urban Environment. Transactions of the Institute of British Geographers. 38(3) 464-479.

Thursday, January 15, 2015

Informational Magnetism on Wikipedia: geographic networks of edits

Previous work that I've published about the geography of contributions to Wikipedia showed the varying types of local engagement that different regions havethe primary reason that Sub-Saharan Africa, in particular, has such a low proportion of locally created content, and some of the ways that Sub-Saharan Africa’s already extremely low proportion of local contributions is inflated by just a few outliers (meaning that the vast majority of content written about countries in Sub-Saharan Africa is created from outside the region).

In all of that work, we employed a simplified distinction of 'within region' or 'outside region' when considering the geography of edits. However, it is possible to also look at which regions send edits to which other regions. Thus, instead of a univariate map or a bivariate scatter plot, we can represent the data as a network of flows.
Screen Shot 2014-12-11 at 22.14.19
The network above is just such an attempt. In it, edge (a line connecting nodes) thickness is proportionate to the total number of edits received. That is, if a region receives most of its edits from North America then the edge from North America to that region will be thick.

When looking at flows of geographic information, there are “net-importing” regions (Asia, Latin America & Caribbean, Middle East & North Africa, and Sub-Saharan Africa) and “net-exporting” regions (North America, Oceania) in terms of edits. Sub-Saharan Africa receives 10.7 more edits from outside than it commits to the outside. North America, in contrast has a figure of 0.38.

The next post will look at the data somewhat differently. Instead of where edits are received from, it will focus on where edits are sent to (as we shall see, this is an important difference).

Note that this work comes from the following report (the relevant parts of which were written in collaboration with Ralph Straumann and Bernie Hogan):

Graham, M., and Hogan, B. 2014. Uneven Openness: Barriers to MENA Representation on Wikipedia. Oxford Internet Institute Report, Oxford UK.

The following paper also offers an abridged version of some of the results:

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).

Or for a broader discussion about why the locality of participation matters, see:

Graham, M. 2014. Internet Geographies: Data Shadows and Digital Divisions of Labour. In Society and the Internet: How Networks of Information and Communication are Changing our Lives, eds M. Graham and W. H. Dutton. Oxford: Oxford University Press. 99-116.

Graham, M., M. Zook, and A. Boulton. 2013. Augmented Reality in the Urban Environment. Transactions of the Institute of British Geographers. 38(3) 464-479.

Thursday, December 4, 2014

Visualising the locality of participation and voice on Wikipedia

On Sunday, I use an extended stopover in San Francisco airport to pop into the Wikimedia Headquarters and chat about uneven geographies of voice and representation on the platform.  In looking through some of the previous work that Ralph Straumann, Bernie Hogan, Ahmed Medhat, and I did on the topic, I noticed a few results that we haven't yet had a chance to blog.

The graph above shows us something very interesting about the locality of participation and voice on Wikipedia.

It investigates the topic by looking at the proportion of within-region-edits to Wikipedia articles.

Each region of the world is assigned a colour (North America - brown; Oceania - dark blue; Europe - light blue; Asia - grey; Latin America - yellow; Middle East - green, and Sub-Saharan Africa - red). The vertical axis focuses on the proportion of edits to articles in the region that come from that region [“received edits”] and the horizontal axis focuses on the proportion of a region’s committed edits that stay within that region [“committed edits”]. (note that The data shows anonymous edits only (+), registered edits only (×) and both edit types combined (●))

On the vertical axis of the figure we can see a clear division between regions that are largely able to define themselves and regions that are largely defined by others. The world regions separate into two distinct groups of three (with Asia in the middle): Sub-Saharan Africa, Middle East & North Africa, Latin America & Caribbean receive comparatively few edits from within their territories (around 25 percent). Europe, Oceania and North America on the other hand receive primarily edits from within (around 75 percent). Asia is edited from within and from outside to almost equal degrees. In other words, there are significant parts of the world in which a majority of content is not locally generated.

Asia, Europe and North America all have 73–80 percent of their committed edits staying within their own region. Interestingly, Sub-Saharan Africa commits just slightly less within-region edits at 67 percent. Oceania, Latin America & Caribbean and especially Middle East & North Africa fall behind with 36–60 percent. In other words, not only does the Middle East & North Africa have a lot of non-locally generated content written about it, many of the edits coming from the region are used to write about other parts of the world.

What does this mean?

• Even when editors from Sub-Saharan Africa spend most of their edits within region, their small numbers mean that most content still comes from elsewhere.

• The global cores of North America and Europe self-represent very effectively by focusing on their own regions.

• Content appears to be very sensitive to feedback loops. A lot of content on an area in one language leads to more content in other languages as translations rather than similar local content.

The global cores focus their editing primarily within their own territories. Large amounts of geospatial content show no sign of deterring people from further contributions and editing: as more content exists, so too do more articles to amend, augment, update and build upon. It is possible that a stock of good content may be an attractive “editing ground” for Wikipedians, whereas a scarcity of content, beneath a certain unknown threshold, may – somewhat paradoxically – demotivate people to fill in the blanks. A relative lack of content may further reinforce perceptions amongst editors that little content equates to a small audience that is not worth writing for.

We expand on all of the ideas in a recent report that we published:

Graham, M. and B. Hogan. 2014. Uneven Openness: Barriers to MENA Representation on Wikipedia. Oxford Internet Institute Report, Oxford, UK.

See also:

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).

And they will also be forthcoming in a new paper (that we're currently working on) that focuses specifically on these issues of self-representation.

Thursday, November 6, 2014

Ashby Prize

I am very happy to announce that I paper that I co-authored with Matthew Zook has been awarded the Ashby Prize. The prize is presented by the journal Environment and Planning A for the most innovative papers published in the journal each year.

The full paper is available at the following link  (one perk of the prize is that the paper gets to become open-access):

Graham, M and M. Zook. 2013. Augmented Realities and Uneven Geographies: Exploring the Geo-linguistic Contours of the Web. Environment and Planning A 45(1) 77-99.
This paper analyzes the digital dimensions of places as represented by online, geocoded references to the economic, social, and political experiences of the city. These digital layers are invisible to the naked eye, but form a central component of the augmentations and mediations of place enabled by hundreds of millions of mobile computing devices and other digital technologies. The analysis highlights how these augmentations of place differ across space and language and highlights both the differences and some of the causal factors behind them. This is performed through a global study of all online content indexed within Google Maps, and more specific analyses of the linguistically and topically segregated layers of information over four selected places. The uneven linguistic geographies that this study reveals undoubtedly influence the many ways in which place is enacted and brought into being. The larger aim of this project is to use these initial mappings of the linguistic contours of the geoweb to push forward a broader debate about how augmented inclusions and exclusions, visibilities and invisibilities will shape the way that places become defined, imagined, and experienced.
We are delighted, honoured, and, honestly, surprised to be awarded a 2014 Ashby Prize. It is a wonderful affirmation of the work we put into the paper and our ongoing research on the geographies of information.

The paper started with fairly simple (yet fundamental) questions about the blurred boundaries between our material and digital lives. Namely, where and under what conditions do these worlds intersect? How do they vary across space and language? And how do we explain these patterns? The paper is interdisciplinary in nature and attempts to draw from, and speak to, Geography, Internet studies, Urban studies, and Communication studies. The piece builds heavily on our earlier work on the hybrid nature of material/digital places (see Graham 2011, 2013 or Zook and Graham 2007a, 2007b): arguing that the lines between digital signifiers and the material signified are becoming increasingly blurred as digital content is tightly enrolled into our spatial experiences.

While rooted in a process tracing back decades to the first satellite imagery and GIS projects (or even millennia if one includes annotations to the material environment such as petroglyphs), the connections between places and information related to them have been made particularly visible and usable over the past ten years. The emergence of crowd-sourced information as well as large professional databases, the ubiquity of all manner of sensors, and the recording of ambient data documenting everyday geographies has created a world that is increasingly augmented by digitally stored, geotagged information. This fact coupled with the ready-availability of interfaces (e.g. Google Maps, Wikipedia, Tripadvisor) through which we can access this information, whilst we’re in the very places being portrayed, has meant a diffusion of the power to create and use spatial representations; an ability once largely limited to a few. In short, digital augmentations are profoundly imbricated into our everyday lives and places.

The augmentations, however, are neither uniform or unidirectional and can both highlight and hide information. Articulating these dual process of uncovering and obscuring is the main goal of this paper, as language is a particularly ready indicator of the unevenness of digital representations. The world, as seen through lens of a search engine or location based service, for instance, simply appears differently depending upon the linguistic contexts through which you engage. The power of maps and spatial representations endures. But the levers through which this power is enacted are now embedded in lines of software, designs of platforms, availability of data sources, and opaque (and often decentralized) governance and production systems, rather than the pen and ink of cartographers or the interests of centralized states or competing elites.

Eventually, information is not only providing the raw material for creating representations of places but actively engaged in augmenting them whilst they are being enacted and brought into being; a theme that we explore in much more detail in a parallel paper on augmented realities we wrote with our colleague Andrew Boulton (Graham, Zook, and Boulton 2013). Today, many people don’t think twice about conducting a spatial search for restaurant or being given bespoke directions home from our precise location. But less than a decade ago, these spatial services were largely limited to GIS labs. We still remember the excitement with which we greeted the first virtual globes and the extreme novelty of a map based search. Moreover, we'll admit to literally spending hours simply entering a range of search terms and locations into Google Local (an early branding of Google Maps) simply to better understand how this new and monumental thing worked and how parts of the world were both revealed and hidden in these representations.

The principle point of all our this work is that the digital layers of places need to be better interrogated. We need to understand what they include and exclude; what they show and don’t show; who does and doesn’t participate in their creation; and what people, platforms and algorithms have a hand in filtering, ranking, and censoring that content. In short, digital dimensions have become key variables for studying all kinds of human geographies, and it is important for researchers to better understand their connections to spaces and places.

Our paper was an attempt to address some of methodological and theoretical issues associated with informational geographies. Using a relatively simple set of software scripts to collect data, we were able to show both that some parts of the world are surrounded by much denser layers of digital information than others, but also the variable nature of this content within places as certain people and languages had considerably more data available to them. In other words, the paper highlighted a number of the balkanised bubbles of augmented information that have important implications for the enaction of material balkanised spaces; a point we take up in later work (Graham, Zook, and Boulton 2013; Zook, Graham and Boulton, 2014). Clearly, these are early days in this body of research and we look forward to many and fruitful exchanges with the growing number of geographers engaged in this work.

As with all academic pursuits, this paper did not emerge from a vacuum and we would like to recognise the broader conversations in which our paper is embedded. The initial draft of this paper was targeted at a 2011 conference held in Oxford by the Oxford Internet Institute -- “A Decade in Internet Time: symposium on the dynamics of the Internet and society – and the interdisciplinary nature of the event helped broaden our vision about who we hoped to engage. That conference paper was subsequently reworked for a special issue of Environment and Planning A on ‘Neogeography’ put together by Matthew Wilson and Mark Graham (Wilson and Graham 2013) which provides valuable context to this work. For example, the special issue also is the home for a selection of papers that address important questions about crowdsourced cartography, the ‘democratisation’ of knowledge production, and power in decentralised networks.

We hope that our paper, and the special issue of Environment and Planning A to which it belongs, become the beginning of a sustained and extensive effort to critically tackle questions about the digital and augmented facets of our lives. But, even as mobile devices and information services become ever more prevalent, our abilities to recognize and research face serious challenges. Many digital layers are opaque, invisible to the naked eye, locked away in closed and ephemeral databases, and tightly controlled by entities that have no interest in critical research. For instance, how and why exactly does Google rank the results of spatial searches? How are these results personalized? How are our updates to social media being captured, processed and repurposed? Are we being subtly directed away from parts of the city based on our social networks and search histories? These are just some of the key questions in this area, and we believe that it is precisely our role as geographers that requires us to constantly attempt to understand how datasets, data intermediaries, systems, platforms, devices, algorithms, and networked crowds work to influence the digital augmentations to our everyday lives and geographies. Maps have power; which makes personalized geovisualizations all the more potent, and understanding the processes behind this, all the more important.

We would like to end with some words from Tony Benn, a British Labour party politician who sadly passed away earlier this year. He famously had a set of five questions that he said that we should always ask any powerful person: "What power have you got? Where did you get it from? In whose interests do you exercise it? To whom are you accountable? And how can we get rid of you?”
Perhaps we should attempt to more explicitly adopt those questions to research into the powerful platforms, algorithms, and systems that govern and structure our networked interactions. We hope that our paper has contributed towards that goal, and we’re excited that so many of our colleagues and collaborators are also working towards those aims.

Mark Graham and Matthew Zook

Graham, M. 2011. Cloud Collaboration: Peer-Production and the Engineering of the Internet. In Engineering Earth. ed. Brunn, S. New York: Springer, 67-83.
Graham, M., M. Zook., and A. Boulton. 2013. Augmented Reality in the Urban Environment: contested content and the duplicity of code. Transactions of the Institute of British Geographers. 38(3), 464-479.
Wilson, M and M. Graham. 2013. Guest Editorial: Situating Neogeography. Environment and Planning A 45(1) 3-9.
Zook, M. and M. Graham. 2007b. Mapping DigiPlace: Geo-coded Internet Data and the Perception of Place. Environment and Planning B. 466-482.
Zook, M., Graham, M, and A. Boulton. 2014. Crowd-Sourced Augmented Realities: Social Media and the Power of Digital Representation. Chapter in S. Mains, J. Cupples, and C. Lukinbeal. Mediated Geographies/Geographies of Media. Springer Science International Handbooks in Human Geography.