ArticlePDF Available

In the name of Development: Power, profit and the datafication of the global South



We examine the current 'datafication' process underway in low- and middle-income countries (LMICs), and the power shifts it is creating in the field of international development. The use of new communications and database technologies in LMICs is generating 'big data' (for example from the use of mobile phones, mobile-based financial services and the internet) which is collected and processed by corporations. When shared, these data are also becoming a potentially valuable resource for development research and policy. With these new sources of data, new power structures are emerging within the field of development. We identify two trends in particular, illustrating them with examples: first, the empowerment of public-private partnerships around datafication in LMICs and the consequently growing agency of corporations as development actors. Second, the way commercially generated big data is becoming the foundation for country-level 'data doubles', i.e. digital representations of social phenomena and/or territories that are created in parallel with, and sometimes in lieu of, national data and statistics. We explore the resulting shift from legibility (Scott, 1998) to visibility, and the implications of seeing development interventions as a byproduct of larger-scale processes of informational capitalism.
In the name of Development: Power, profit and the datafication
of the global South
Linnet Taylor
, Dennis Broeders
University of Amsterdam, The Netherlands
Erasmus University Rotterdam, The Netherlands
article info
Article history:
Received 15 October 2014
Received in revised form 2 July 2015
Accepted 3 July 2015
Available online 9 July 2015
Big data
Public private partnerships
We examine the current ‘datafication’ process underway in low- and middle-income countries (LMICs),
and the power shifts it is creating in the field of international development. The use of new communica-
tions and database technologies in LMICs is generating ‘big data’ (for example from the use of mobile
phones, mobile-based financial services and the internet) which is collected and processed by corpora-
tions. When shared, these data are also becoming a potentially valuable resource for development
research and policy. With these new sources of data, new power structures are emerging within the field
of development. We identify two trends in particular, illustrating them with examples: first, the empow-
erment of public–private partnerships around datafication in LMICs and the consequently growing
agency of corporations as development actors. Second, the way commercially generated big data is
becoming the foundation for country-level ‘data doubles’, i.e. digital representations of social phenomena
and/or territories that are created in parallel with, and sometimes in lieu of, national data and statistics.
We explore the resulting shift from legibility (Scott, 1998) to visibility, and the implications of seeing
development interventions as a byproduct of larger-scale processes of informational capitalism.
Ó2015 Elsevier Ltd. All rights reserved.
1. Introduction
There is a process of ‘datafication’ (Mayer-Schönberger and
Cukier, 2013) underway in low- and middle-income countries
where the use of new communications and database tech-
nologies in LMICs is generating digital data that is machine-readable
and computationally manipulable, particularly for ‘big data’ analyt-
ics. These born-digital datasets are of unprecedented size and detail,
especially compared to the statistical records previously available on
lower-income countries (Jerven, 2013). In contrast to traditional
state survey data, however, these data are generated, collected and
processed under the auspices of private-sector corporations and
are shared, often on a pro-bono basis, at the level of international
academic research institutions or development actors such as the
UN. Where previously development donors (governments or
international NGOs) worked with LMICs’ own statistical apparatuses
to generate population data, it is becoming increasingly possible
and cost-efficient for donors to turn to corporations for
consumer-generated data that can proxy for traditional household
surveys and other statistical products (Taylor and Schroeder,
2014). The discourse on big data as a resource for development
(World Economic Forum, 2012; Global Pulse, 2012; Taylor and
Schroeder, 2014) indicates that a shift is underway from the pre-
dominance of state-collected data as a way of defining identities
and sorting and categorising individuals, groups and whole societies
to a big-data model where data is primarily collected and processed
by corporations and only secondarily accessed by governmental
The central question addressed in this paper is how datafication
is influencing the way that LMIC populations are made legible in
the context of development, and what this means for power
dynamics amongst development actors. We examine a power shift
from the traditional collector and user of statistics – the state – to a
messier, more distributed landscape of governance where power
accrues to those who hold the most data. This power shift has its
roots in the larger neoliberal trend in governance worldwide –
what Cohen (2013: 1928) has termed the ‘new governance,’ dom-
inated by public–private partnerships engaged in ‘informational
capitalism’ (2013: 1912, following Castells (1996)), a system where
‘information flows in circuits that serve the interests of powerful
0016-7185/Ó2015 Elsevier Ltd. All rights reserved.
Corresponding author.
E-mail address: (L. Taylor).
We use the World Bank’s definitions grouping countries, see: http ://data., where LMICs have incomes of
US$1,036 – $12,616 per capita and high income countries (HICS) above that
threshold. Our particular focus is the low- and lower-middle-income countries, with
an upper threshold of $4,085 per capita, which includes India and most of Africa.
Geoforum 64 (2015) 229–237
Contents lists available at ScienceDirect
journal homepage:
entities, both private and public’ (2013: 1916) and where those
who trade in information – primarily corporations – are able to
modulate people’s behaviour, activities, and relationship to the
state. She argues that the creation of surveillance infrastructures
are an inevitable part of governance through informationalism,
and the possibility of surveillance is therefore pervasive. We argue
that ‘data-driven development’ is characterised by the same fea-
tures as informational capitalism, and results in increased visibility
for the populations of lower-income countries – though not neces-
sarily in greater governability or representation.
This shift towards a combination of datafication and privatisa-
tion is still in its early stages and the evidence is not yet available
to draw conclusions about its medium or longer-term impacts. It is
potentially wide-ranging, however, and has powerful implications
for how LMICs will be understood by the development field as eco-
nomic and political actors in the coming years. So far, little
research has been conducted on the political implications of the
production of big data about lower-income countries (exceptions
include Thomas (2014), Burns (2014) and Taylor and Schroeder
(2014)). We therefore assess the longer-term risks of the secondary
use of such data under the rubric of economic or human develop-
ment, based on illustrative examples of current data collection by
corporations and public–private partnerships in LMICs and aim
to contribute towards a research agenda on this issue by drawing
out and illustrating key trends and their potential repercussions.
We define two key trends using illustrative examples: first, the
increasing empowerment of private sector actors in the field of
international development due to their ownership of data, and sec-
ond, the emergence of development/surveillance assemblages with
the potential for ongoing monitoring of population dynamics and
people’s activities. The first trend involves leverage and opportuni-
ties for corporate actors who are proprietors of vast quantities of
data on citizens of developing countries, whilst states continue to
prioritise traditional survey data such as censuses which convey
a different type of detail and allow for different types of sorting
and monitoring. We look at how the deputisation of technology
corporations by states through public–private partnerships (PPPs)
may lead to the merging of these different analytical perspectives.
We then frame the second trend, the emergence of country data
doubles and various forms of shadow mapping – in relation to
these partnerships.
This paper is based on 60 interviews on the use of big data in
development policy and planning, along with two years’ research
(2013–2015) on the interface of data science and international
development policy. The research also involved attending interna-
tional conferences, workshops and public discussions hosted by
large institutional actors, and by gathering knowledge from mail-
ing lists and other online discussions. This project was partly
funded by the Alfred P. Sloan Foundation at the Oxford Internet
Institute. The interviews on big data in the field of international
development were conducted with academic researchers and
private-sector data scientists working with big data on questions
relevant to LMICs. Interviewees were selected using a purposive
sampling process focused both on the most relevant projects, and
on those with an overview of the state of the art in big data
2. Power through LMIC data
2.1. Legibility, visibility and emergent power structures
Scott’s (1998) famous book Seeing Like a State could instead
have been titled ‘reading like a state’ since the registrations upon
which the modern bureaucratic state was built, made society –
in Scott’s own words – ‘legible’. ‘Reading like a state’ in the era of
big data, however, increasingly involves remotely performing data
analytics to make populations visible. This constitutes a step
beyond – or away from – Scott’s formulation because such visibil-
ity creates power over data subjects via volume of data rather than
accuracy and detail. Unlike legibility, which depends on data that
people are aware of providing, the data that provides visibility is
often observed, not volunteered (Hildebrandt, 2013), derived as a
byproduct of technology use rather than collected by authorities
through survey methods. Unlike the data that provides legibility,
these data are of unknown reliability because they reflect, in
Shearmur’s words (2015), not populations but ‘users and markets’
and are therefore biased towards those with connectivity. This
unknown bias is compounded because the visibility such data
offers is created by data scientists, not social scientists, and fre-
quently lacks contextual information to explain what is being seen
(Taylor and Schroeder, 2014). The gap in terms of distance – but
also cultural dissimilarity - between the processors and end users
of LMIC big data and those living in these countries may also
increase power asymmetries. This suggests that the power to make
visible is different from the power to make legible. Legibility
increases governability (in Scott’s formulation), but visibility offers
the power to influence and intervene to a wider, more distributed
set of actors: the corporations who gather and analyse the data,
plus whoever they choose to share it with (or can capture it
through other means), who may be state actors, international
development institutions, or other corporate partners.
Traditional visualisations such as mapping are being trans-
formed by spatial big data analytics. GPS-derived big data from
mobile phones in particular lends itself to real-time, highly
detailed mapping of mobilities (Taylor, 2015a), and sensor data
offer new possibilities for understanding urban space in particular
(Kitchin, 2014). However, this big data is mainly created and
accessed within government–corporate networks (Crampton
et al., 2014), so that big data is a tool of the powerful in the corpo-
rate and governmental spheres who can afford to collect, merge
and analyse it. In this way, little has changed since Harley’s
assertion (1989) that maps are essentially political despite the
claim of scientific objectivity put forward by the profession of car-
tographers. It also reinforces and broadens Taylor and Johnston’s
(1995: 58) warning to geographers that by using digital data gen-
erated by state apparatuses, they are ‘creating the state’s geogra-
phy’ and determining ‘what is included and what is not’. Today, a
similarly critical approach is needed in relation to corporations
that collect and own data, and determine how they may be made
accessible. What kind of geography is emerging though the
corporate-led Data for Development?
Power is thus exerted through the generation of data, control of
access to data, the merging of various data, but also through their
analysis and interpretation. At each of these points, however, the
private sector’s role is increasingly important but also increasingly
hard to pin down. Bowker and Star (2000) in their seminal work on
classification remarked that information architectures (and the
politics that they contain) are hard to analyse because ‘Good usable
systems disappear almost by definition. The easier they are to use,
the harder they are to see. As well, most of the time, the bigger they
are, the harder they are to see’ (Bowker and Star, 2000: 33). At this
moment, big data analytics constitute the biggest usable system
around and the corporate and governmental enthusiasm for its
potential does not motivate many to analyse what lies under the
hood. Similarly, Schneier (2015: 30) points out that data collection
in HICs has become so engrained in modern life that it fades into
the background for the individual user. Moreover, most of the
attention that has been paid to projects under the rubric of Data
for Development focuses on the perceived and expected benefits
at the expense of the epistemological and power implications of
the data analytics used.
230 L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237
For these reasons, data analytics are becoming the new battle-
ground of representation. Ultimately, all data lead to actionable
knowledge (Degli Espositi, 2014: 212), or in the words of Amoore
(2011: 27): ‘the processes of data integration, mining and analytics
draw into association an amalgam of disaggregated data, inferring
across the gaps to derive a lively and alert new form of data deriva-
tive – a flag, map or score that will go on to live and act in the world’.
Yet data and representation ultimately serve their ‘owners’ and are
at least in part built on power asymmetries. Graham (2009) writing
about ‘virtual earths’ and web 2.0 applications within webglobes
such as Google Earth and Microsoft Virtual Earth contends that
these are not simply mirrors of their physical counterparts but
are characterised by ‘information black holes’ as well as ‘hubs of
rich description and detail’. He warns that: ‘Although myriad repre-
sentations of place are now easily accessible, it remains important
to note that the virtual earth remains highly shaped by dominant
power structures, software algorithms and the cultural links
between producers of information’ (Graham, 2009: 432). Being
seen and not being seen, looking and not looking are in his view part
of a power structure that is almost always asymmetrical. Visual
equality is a rare thing, especially when states are involved.
(In)visibility is therefore a power struggle: ‘Invisibility is a relation-
ship between those who have the power to see or to choose not to
see, and, on the other hand, those who lack the power to demand to
be seen or to protect themselves from the negative effects of
imposed visibility’ (Polzer and Hammond, 2008: 421).
2.2. Distributed governance: PPPs for data and development
If having the data to see what was previously invisible, or
ill-defined, is a source of power, then this power seems to be shift-
ing to a new group of actors. Worldwide, big data is largely propri-
etary, being generated and analysed by technology-providing
corporations. Now that data analytics are being advocated as a
new solution to development problems, those with the data will
also increasingly gain power over how development interventions
are conceptualised and practiced. As corporations acquire the
power to make data subjects visible, they also acquire the power
to monitor (or to ‘aid surveillance’, as Privacy International has
put it (Hosein and Nyst, 2013)). Lyon (2007) argues that surveil-
lance is always somewhere of the continuum between ‘care’ and
‘control’, with the latter more closely associated with the (nega-
tive) political label. In the world of Development, monitoring
through data is constant and widespread, and the power to see is
also the power to influence: Lyon (2007) defines surveillance as
‘(...) the focused, systematic and routine attention to personal
details for purposes of influence, management, protection or direc-
tion’. Organising aid, development interventions or even welfare
states require data for monitoring inclusion, progress and impact
– but in the development field, where there is an accountability
gap with regard to international interventions, it is important to
understand how data may also enable the abuse of power.
Jerven (2013) shows that statistical apparatuses in
lower-income countries are widely judged to be inadequate and
under-resourced, a problem which has recently led researchers
and policymakers to identify ways in which new sources of digital
data, arising from the use of digital technologies by people in
LMICs, may provide a way to supplement traditional
survey-based methods of data-gathering such as censuses (e.g.
Center for Global Development, 2014; Green, 2014; Global Pulse,
2013). This high-level institutional interest in ‘data-driven devel-
opment’ (World Economic Forum, 2012) has been accompanied
by the notion of ‘data philanthropy’ (Kirkpatrick, 2011), where
the corporations providing digital services to populations of
LMICs give development authorities access to that data under the
rubric of Corporate Social Responsibility (CSR). This process
involves high-level intermediaries, notably UN Global Pulse, one
of the highest-profile nodes in an informal network of interna-
tional researchers and development organisations dealing with
data-driven development projects, publishing and discussing
results, and sharing the knowledge gained from interventions.
Examples of past projects within this network include the 2013
Data for Development challenge established by Orange (2012),
the donation of mobile phone data to epidemiologists to track
the cholera outbreak after the 2010 Haiti earthquake (Bengtsson
et al., 2011), and UN Global Pulse’s use of social media to track food
price shocks in Indonesia (Global Pulse, 2014). Each of these
research projects has been conducted with ‘big data’ donated free
of charge by large multinational corporations such as Orange and
Twitter for purposes of monitoring and evaluating social or eco-
nomic dynamics in LMICs.
The discourse on big data as a resource for development (World
Economic Forum, 2012; Global Pulse, 2012; Taylor and Schroeder,
2014) highlights that data is primarily collected and processed by
corporations and only secondarily accessed by governmental
authorities. This shift has two effects: first, it translates the individ-
ual from citizen to data subject, ‘a conditional form of existence
whose rights are dependent upon its behaviour within digital net-
works’ (Bauman et al., 2014: 129). Second, it underlines the move
to a more distributed governance model with regard to population
data, ‘organized neither horizontally, in the manner of an interna-
tionalized array of more or less self-determining and territorialized
states, nor vertically in the manner of a hierarchy of higher and lower
authorities.’ (ibid: 124). Instead, those who hold the data increas-
ingly have the power to intervene, or to inform interventions.
Decentralised or distributed governance may come about due to
conditions of limited statehood (Livingstone and Walter-Drop,
2014), but may also arise through the involvement of international
development institutions, including private-sector actors. The
legacy of the household survey sponsored by large development
institutions (e.g. USAID’s Demographic and Health Surveys; the
InDepth Network’s health and demographic surveillance system,
and UNICEF’s Multiple Indicator Cluster Surveys), along with a
myriad country-level examples run by Non-Governmental
Organisations (NGOs), have created a landscape where data flows
in complex circuits which sometimes intersect with the host state’s
own, and sometimes do not. Thus foreign institutions,
non-governmental organisations and public–private partnerships
may all play a role in collecting data to make populations visible,
and in creating and monitoring interventions in social and eco-
nomic life.
Another repercussion of this shift is that the power to profile is
in new hands. According to Lyon (2014: 6–9) the use of big data
results in the automation of analysis, which gives those who
develop algorithms for sorting and classifying data subjects signif-
icant power over those people’s lives – more so if the analysis is
being used to direct development interventions. In the era of digi-
tal communications and big data this distributed governance of the
means of sorting and monitoring populations is translated into
unrestrained digital data maximisation and, in turn, to what
Haggerty and Ericson (2000) have termed ‘data doubles’ –
abstracted representations of people through their data, formed
with the aim of monitoring or targeting people for intervention.
Where corporations are responsible for profiling, the data doubles
that are created are a source of power in themselves. Ruppert
(2012) has updated this concept for the era of big data, noting that
data doubles are now constructed by a plethora of different data
processors, and Lyon (2014) adds that big data analytics give the
data double more power (and its subject greater visibility) by mak-
ing it ubiquitous and constantly self-generating. In line with these
contributions we scale up the notion of the data double, showing
how it now extends to the collective level of cities and countries.
L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237 231
Where corporations conduct or facilitate development research on
LMICs, it is worth examining the risks of repurposing data analytic
techniques developed for purposes of profit-making. As Lyon notes,
‘The enthusiasm for big data ‘‘solutions’’ may lead to inappropriate
transfer of techniques from one field to another’ (Lyon, 2014: 6).
Given all these factors, we therefore draw attention to a shift in
power relating to LMIC development: namely, that the power to
count, categorise and visualise those in LMICs is increasingly being
acquired by corporations, particularly through the use of new com-
munications technologies by LMIC citizens. This shift is towards
practices of informational capitalism, which has as its byproduct
a new level of visibility for previously less-surveyed populations,
or at least those in contact with digital technologies. We posit that
this shift extends unprecedented power to monitor, and therefore
also to surveil, to new and unregulated actors; that it gives corpo-
rations a new type of influence over the lives of individuals with
potential for misuse, and that it may exclude smaller-scale actors
with local knowledge and understanding. In what follows, we offer
illustrations of this shift and draw out the implications for its
longer-term repercussions.
3. Two trends in datafication and development
In this section we explore two trends which illustrate the issues
identified above. We first show how corporations’ technical capac-
ity leads to their deputisation by LMIC states and foreign donors
through the institutional mechanism of Public–Private
Partnerships (PPPs), and how such projects are using the develop-
ment discourses that underpin these partnerships to justify the
search for new markets. Secondly, we examine how private big
data are becoming the foundation for national-scale data doubles
and shadow maps, i.e. digital representations of social phenomena
and/or territories that are created in parallel to, and sometimes in
lieu of, national data and statistics. These PPPs are thus leading to
new data doubles and mapping systems which cannot be sepa-
rated from the power structures and political economy that create
them, and that they reproduce through their manipulation of data.
3.1. Corporations as development actors, development as market
As corporations expand into emerging markets through services
which generate digital data, they now find themselves simultane-
ously expanding into the development field. Although the private
sector has long played a role in development through Corporate
Social Responsibility (CSR) activities, corporations are becoming
involved in the field in new ways as digital data emerges as a
resource for analysing development and informing policy. This is
taking place in two different configurations: first, with corpora-
tions as partners or deputies of states or donors where the latter
are still developing digital capacity, a configuration which reflects
the distributed modes of governance facilitated by the new data
flows. Second, there are projects which carry the label of develop-
ment but which are managed and rolled out entirely by corporate
interests. This section draws on three different projects to examine
how discourses of development are mingling with those of profit as
corporations increasingly move into the space previously occupied
by states and nonprofit donor institutions.
An example of the first configuration can be seen in India’s bio-
metric identification project, the Unique Identification Authority of
India (UIDAI, also known as Aadhaar), the largest-scale public–pri-
vate partnership currently underway in a developing country in
terms of its coverage of the population. This scheme, in operation
since 2008, forms the world’s largest biometric database and aims
to provide every Indian resident with a 12-digit Unique Identity
number (, 2014). By mid-2013, 400 million people’s
details had been entered (Business Standard News, 2013). The pub-
lic–private structure of the scheme was inevitable due to the
amount of technical expertise necessary to deal with the big data
analytics that are essential to real-time crosschecking and updat-
ing of a database of hundreds of millions of records (Krishnan,
2012). Besides acting as identification for welfare recipients, a
UID is also the gateway to a bank account, a mobile phone contract,
and various commercial products which corporations previously
were unable to offer to lower-income citizens partly because of a
lack of reliable identification. The project, however, is only dubi-
ously ‘public’. Officially UIDAI is a quasi-governmental organisa-
tion attached to the national planning authority, but effectively it
is an autonomous organisation, operating without a legal frame-
work or parliamentary oversight (Ramanathan, 2013). Besides
being non-governmental, UIDAI is also a collaboration between
and amongst corporations. It is managed and chaired by Nandan
Nilekani, a co-founder of InfoSys, one of India’s largest technology
and consulting firms, and the day-to-day work of gathering, pro-
cessing and storing data is done by private companies (India
Today, 2013; Yadav, 2013). The collaboration bears out Miraftab’s
claim (2004) that corporate involvement in the Development pro-
ject is conceptualised as a way of importing technical and manage-
rial capacity to under-resourced countries; but as he points out,
since the oversight role ascribed to government authorities in
those countries is often only sketchily established, corporations
often remain essentially autonomous actors and these arrange-
ments may come to underpin distributed governance structures.
The UIDAI project is extremely broad in its aims: it is presented
by Nilekani as facilitating development and poverty alleviation; ‘a
huge project of social inclusion ...[that will] improve the quality of
government spending on various public programs’, and a way of
becoming visible to the ‘welfare state’ (Nilekani, 2013). He also
states, however, that the project could be a blueprint for ‘a
next-generation healthcare record system for developing coun-
tries, or you could use it for massive online education, or what-
ever.’ (ibid). The other functions constituted by ‘whatever’ range
from giving marketers direct access to potential consumers to cre-
ating linkable data on individuals’ location, consumption and other
characteristics. These are hugely commercially powerful, and pos-
sibly eclipse the ID’s role as a connection to the state’s welfare
apparatus. As Nilekani describes it, the database is designed to be
a complement to state records, and can therefore enable a variety
of consumer functions as much as operations of citizenship.
Nilekani’s speech to the Center for Global Development (ibid) also
includes a vision of the UID scheme as a form of instant direct mar-
keting on an unprecedented scale: ‘The network effect also comes
into play. Let’s say that we have 300 million people who have this
ID, and let’s say that somebody launches a product for a new kind
of an annuity product. With the help of our system, all 300 million
people can instantly get access to the product. So, it really creates
this huge network effect where you can quickly get people to ramp
up and get services.’ (Nilekani, 2013). Nilekani has subsequently
suggested that the UID database will be opened up to app develop-
ers in order to speed the invention of ways to connect corporations
directly to customers (LiveMint, 2013). These are the features
which have led Jayaram (2014) to refer to the scheme as a ‘welfare
industrial complex’ with ‘implications for how citizens relate to
private sector entities, on which the UID rests and which have their
own vested interests in the data’.
As Nilekani’s speech points out, the UID was never designed as
just a public development project, but has a commercial rationale
at its core. This combination of functions both extends the UID pro-
ject into the realm of governance and raises serious questions
about privacy (Jayaram, 2014). This both aligns with Cohen’s vision
of informational capitalism (Cohen, 2013), and is also an example
232 L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237
of data gathering moving from objectives of care towards a reality
of control (Lyon, 2007). As Dreze (2010) has pointed out, the link-
ing of the UID to the National Population Register is also poten-
tially a way of relating individual biometric ID numbers to tax
returns, bank records, SIM (subscriber identity module) registers
and the Indian Home Ministry’s National Intelligence Grid
(NATGRID), which in turn links to 21 national databases. In fact,
Dreze’s predictions have come true as domestic terrorism has
become more of an issue in India over the six years of the UID’s
existence. The scheme’s security implications have become an
important selling point – demonstrating how population-level
information systems easily become subject to function creep –
the biometric database has been planned to be compatible with
‘national security and border control applications’ (Zelazny, 2012).
Alternatively, corporations may choose autonomously to carry
out work they define as development-related, rather than aiming
to fill a gap left by state capacity. This can be seen occurring in
important emerging markets such as East Africa, as with the recent
rollout of IBM’s Project Lucy in Kenya. IBM promises to ‘solve
Africa’s grand challenges’ including ‘healthcare, education, water
and sanitation, human mobility and agriculture’ (IBM, 2014) using
artificial intelligence and big data analytics. The project will take
place within a private–public consortium structure led and man-
aged by IBM, including ‘government leaders, university professors,
business leaders, heads of nonprofits and people involved in the
burgeoning tech startup communities’ (IBM, 2014: 4). It will
involve feeding all the published economic and social data avail-
able from Sub-Saharan African countries into IBM’s ‘Watson’
supercomputer, which will then data-mine for answers to ques-
tions. (So far the company is not responding to researchers’
requests for more information about Project Lucy, and the materi-
als available do not specify whether the project will also have
access to data which is not in the public domain.) Despite the pro-
mise to address problems which have their roots in poverty, the
lack of a strong civil society, inequality and inadequate governance,
the leader of IBM’s Project Lucy is clear that the standard for suc-
cess is profitability, saying that ‘We’re creating a strong
business-oriented culture in the lab... I want my people to con-
stantly ask themselves, ‘‘Who’s going to buy it?’’’ (IBM, 2014: 7).
Similarly, the chief data scientist for the project again blends the
discourses of development and corporate profit when he warns
that ‘You have to be aware of the price point (for Project Lucy’s
findings). You have to be aware of cultural issues, of education.
You have to make technology easily consumable.’ (IBM, 2014: 4).
Project Lucy epitomises how corporations’ power over data trans-
lates into power to determine what constitutes development. Any
organisation that can pay may ask questions of the Lucy database,
but the structure and content of that database has been deter-
mined by IBM – a company whose motives in the region are pri-
marily profit and market share.
Perhaps the most high-profile project in the corporate/develop-
ment marketplace currently is that of Facebook’s pro-
ject, which promises to offer the internet to those previously
excluded by poverty. As of 2013 69 per cent of the world’s mobile
subscribers had only 2G service (GSMA, 2013b), suggesting that
brand recognition and subscribers captured whilst connections
are slow will yield an important market advantage as they speed
up over time. Facebook’s aims to capture this market
by extending internet connectivity, primarily via mobile phones,
‘to the two thirds of the world’s population that doesn’t have it’
(, 2014). The project will use ‘drones, satellites and
lasers to deliver the internet to everyone’.
In combination with
providing connectivity, Facebook also promises mobile subscribers
in LMICs free data service on their phones when they access a
text-only version of Facebook and selected ‘partner sites’,
for other sites users will have to pay data charges. Facebook markets
this ’Facebook Zero’ service as a way to enable low-income people in
areas with low levels of technological infrastructure to go online in a
simpler internet environment.
However, despite Facebook’s rhetoric about the benefits of ‘con-
necting the world’ to the internet (see note 3), they are instead
connecting a subset of the world. The Facebook Zero service is
being marketed to the most populous emerging economies, where
technology corporations are looking to build market share: initially
Indonesia, the Philippines, and Malaysia, and soon thereafter
Nigeria and South Africa (MTN, 2010). In terms of the global
South, then, the project focuses mainly on those countries on the
verge of connecting themselves, where awareness of the internet
is high, and desire for online social networking so powerful that
the mobile phone becomes just a vehicle for a particular applica-
tion. Therefore reducing the asymmetries in connectivity between
more profitable markets and remoter, lower-income areas will not
be an outcome of the current strategy, although the company’s
rhetoric strongly implies it. Instead the project is likely to reduce
the asymmetry in the use of Facebook. An activist in India inter-
viewed for this project said of Facebook and other social networks:
That’s being used to sell phones now. I mean, I remember doing
this at one point, just taking lots of photographs of hoardings that
were advertising phones, and they’re not advertising the phone,
the phone is a very small part in the corner of the billboard, it’s
Facebook or Twitter which is... you know, it’s connected to Find
Your Friends, that’s how the phone is being sold now.
[(Digital rights activist, India, interviewed 12.11.2013)]
Providing internet connectivity based on access to Facebook,
then, is likely to convince first-time users that they are accessing
the most important services available, i.e. social networking. In
doing so, the project reduces the internet to what the Electronic
Frontier Foundation (EFF, 2015) has called a ‘walled garden’ of
tethered partner services which restricts the possibility for creative
use of the web, and for making it impossible for users to use secure
connections. In the words of the Electronic Frontier Foundation,
‘ is not neutral, not secure, and not the internet’ (EFF,
Facebook’s strategy is logical in a saturated market: the next
billion people online will be in the countries focuses
on. The project, however, does more than capture new markets.
It underlines how corporate projects publicised as promoting
development and inclusion may still reinforce existing power
and knowledge asymmetries, and it highlights how the data corpo-
rations gather from developing countries is likely to be skewed by
particular business strategies, with consequences for whether and
how donor institutions should align with these corporate actors.
The project will also have a less obvious but possibly more insidi-
ous effect, however. Facebook’s connection to millions of previ-
ously unreachable subscribers will result in new datasets on
their characteristics and behaviour, and consequently new types
of visibility for those people. If the power to visualise is also the
power to influence and intervene, Facebook’s project empowers
Facebook first, but next its strategic allies such as app developers
and advertisers, creating new axes of visibility and power as a
byproduct of the aim to create new digital consumers.
Informational capitalism, in this case, will create both visibilities
and the power to represent. In the process it will also, by capturing
Statement via Facebook by Mark Zuckerberg, 27.3.2014: https://www.face-
Statement via Facebook by Sid Murlidhar, 18.5.2010:
L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237 233
data doubles in the global South, distribute the power to govern to
anyone who can afford it.
The examples above suggest that corporations’ data-gathering
and analytical advantage means they no longer have to conceal
for-profit motives under development discourse, as was the case
with CSR projects, but instead are now free to connect the expan-
sion of markets, and the attendant expansion of digital records,
with human development. This perspective is being answered by
an increasingly technocratic and commercial discourse amongst
the new coordinating bodies such as UN Global Pulse and the
Stanford Peace Innovation Lab. Like corporations, they use the lan-
guage of the private sector to frame human and economic develop-
ment aims. One example comes from the director of the Peace
Innovation Lab at Stanford:
‘If you can measure something, you can design for it; if you can
design for it you can create new value; if you can create new
value you can monetize it. Our aim is to create peace
Similarly, the methodologies used by Global Pulse’s analytics
lab are common to marketing researchers and business analysts,
and the lab’s director has said that Global Pulse is ‘trying to track
unemployment and disease as if it were a brand’.
The rise of the ‘welfare–industrial complex’ (Jayaram, 2014)
combined with the availability of vast new digital datasets has
increased the risk that human development will become framed
as an engineering problem.
Cherlet (2014) notes that this engineer-
ing approach to development is based on both classic technological
determinism and a new form of ‘epistemic determinism’, the latter
denoting an approach which does not distinguish between technical
knowledge produced remotely and that based on local understand-
ing. This determinism tends to distance human development priori-
ties from local concerns and understandings of development
problems, and from academic research domains which might add a
historical or political perspective. For example, the ‘grand challenges’
the leaders of Project Lucy propose to solve, including disease, illit-
eracy, contaminated water and food insecurity can also be seen as
grand challenges for politics, governance and activism since they
are strongly related to embedded social and economic inequality,
governance priorities and, often, international power asymmetries
around issues such as trade and debt. Framing these as data and
engineering problems has two repercussions. First, it shuts out the
expertise and knowledge of small-scale organisations, qualitative
researchers and local activists, who have traditionally acted as the
extension agents of states and development organisations in actually
applying pro-poor policies in LMICs; and second, it may divert
power, money and resources from these smaller-scale actors.
Meanwhile, the resources and institutional power currently being
marshalled to support the engineering approach makes for an exten-
sive and in-depth way of monitoring activities and trends as a whole,
making this kind of collaboration arguably more apt as a surveillance
infrastructure than as a way of solving development problems.
3.2. Shadow maps and state data doubles
States have always gathered data on their own populations, and
more recently those with less resources have been funded to do so
by international donors such as USAID (organiser of the worldwide
Demographic and Health Surveys [DHS] program) or UNICEF
(which carries out the Multiple Indicator Cluster Surveys [MICS]).
The sheer cost of building large, longitudinal datasets on developing
countries, however, has led the Center for Global Development, an
influential development think tank, to suggest that development
donors should turn to the private sector and ‘experiment with
new institutional models, such as public–private partnerships or
crowdsourcing, to collect hard-to-obtain data or outsource data col-
lection activities’ (Center for Global Development, 2014). This can
already be seen occurring through the growing visibility of institu-
tions such as UN Global Pulse and corporate data-gathering efforts
such as Project Lucy. States, rather than generating statistics, must
now find ways to access them from the private sector.
And as the new sources of digital data start to parallel or even
supplant national data collection efforts, they will increasingly offer
the opportunity to re-visualise countries and populations – another
element of the gradual shift from legibility towards visibility. As
observed data at the population level becomes a possibility, big data
analytics performed on LMICs are starting to generate state-level
versions of the ‘data doubles’ theorised by Haggerty and Ericson
(2000). The examples above, particularly that of Project Lucy,
denote a process where instead of human bodies becoming
abstracted, whole cities and countries can be datafied, analysed
and interpreted largely in isolation from national processes of data
gathering and analysis. Grumbach and Frénot (2013) have written
in Le Monde that Google now knows more about France than its
national statistical agency, INSEE. If this is the case, then the amount
of new information big data can provide about a country such as
Angola, which did not conduct a census between 1970 and 2014,
is exponentially greater (though also less representative, since a
smaller share of Angola’s population uses digital technologies).
Orange’s 2013 Data for Development project demonstrates two
potential problems with the new country data doubles. The first
‘D4D’ challenge, organised in 2012–13, involved the first major
release of mobile phone calling records from an African country
for research purposes. It was designed to encourage both basic
and applied research using anonymised Call Detail Records (CDR)
from the company’s subscribers in Côte d’Ivoire. The aim of the
challenge was to ‘help address the questions regarding develop-
ment in novel ways’ (Orange, 2012). The company received much
positive publicity in its core (European) markets, and aimed to
use the research findings to plan development interventions in
Côte d’Ivoire. The first iteration of the project resulted in 74
research papers, dealing with human mobility, social and economic
development (mapping poverty, tracking economic and social
activity), data mining and health (Netmob, 2013).
The Côte d’Ivoire project also demonstrates how the visibilities
produced by big data analytics (the research findings were
expressed predominantly through visualisations and network
analysis) may be at the same time rhetorically powerful and yet
unreliable due to their unknown bias. They illustrate Shearmur’s
point (2015) that such datasets reflect users and markets, with
one research team (Mao et al., 2013) pointing out that their find-
ings were based on ‘a significant sample of all Côte d’Ivoire mobile
users that we assume is not strongly biased, lacking data to the con-
trary’ (italics added). Gutierrez et al. (unpublished), similarly
warned of the ‘unreliability of census reports’ from Côte d’Ivoire
provided as ground truth for the remote research performed. The
analysis by Berlingerio et al. (2013), for example, shows the effects
of a lack of supplementary data to give context to digital traces.
Creating a transport optimisation model for Abidjan, Côte
d’Ivoire’s capital city, the researchers were provided with a dataset
that offered only a partial image of the transport network that did
not show what factors might be impeding traffic flows (Taylor,
2015b). Nevertheless, the data double assembled from the avail-
able (online) sources was used to create a model which was then
Mark Nelson, Stanford Peace Innovation Lab Co-Director. Panel at the Leiden
University Big Data for Peace Summer School, August 16–22, The Hague.
Robert Kirkpatrick, director, UN Global Pulse, panel at Leiden Innovation Lab,
The discourse on development as a unitary issue lending itself to engineering
solutions is common to many big data projects, for example the ‘Big Data for Social
Good’ project at Harvard:
234 L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237
put forward as a finished data assemblage, ready to guide transport
policy. Although the model was designed to apply to Côte d’Ivoire’s
whole transport sector, the data in fact reflected only the parts of it
that were expressed through digital data. This mismatch shows on
the one hand how big data can mislead through bias and lack of
context, but on the other how corporations that hold data can be
empowered to draw conclusions on a policy level about developing
countries, despite the potential inaccuracy of the data doubles they
can create.
The new data sources allow not only country-level data dou-
bles, but what might be termed shadow-maps: the use of spatial
data to empower new forms of visibility separate from state map-
ping efforts. For example the new maps being derived at the time
of writing through data donated by mobile phone corporations to
track and predict the Ebola epidemic in West Africa (Talbot,
2014) may have huge implications for foreign donors who wish
to target interventions and funding, but also for those who wish
to enforce quarantine and close borders. The new shadow maps
are set to become ever-more detailed and ubiquitous. Manyika
et al. (2011) have projected that by 2020 LMICs will be the primary
producers of geolocated data from the use of digital communica-
tions technologies such as mobile phone use, social media such
as geolocated Tweets, and the traceability of IP addresses when
individuals use the internet. These are being complemented by
the use of powerful imaging technologies in LMICs including satel-
lites and drones, with the latter becoming an important resource
for commercial, development and humanitarian purposes. Whilst
satellites provide the data used in most types of conventional
maps, the production of spatial data is an inevitable by-product
of drones’ primary functions, which are often commercial. Drones
are advocated as a way of gaining access to rural and remote cus-
tomers in LMICs whilst saving those countries the need to build
roads (BBC News, 2012); they are already deployed by the UN in
peacekeeping operations in the Democratic Republic of Congo
(Crowe, 2013) and by entrepreneurs for humanitarian response
after natural disasters (Churchill, 2014). The World Bank is seeking
to demonstrate the potential of drones for predictive and planning
purposes ‘in many sectors including: cadastral mapping/registra-
tion, infrastructure projects (roads, energy and dams), urban plan-
ning, and disaster risk management’ (Volkmann, 2014).
Each of these LMIC spatial data projects has a different purpose,
is made up of different institutional configurations (private sector
firms such as Orange, multilaterals such as Global Pulse, or
NGO/multilateral collaborations such as that between the World
Bank and various partners in Haiti,
) and is subject to different
forms of governance. However, together they are generating shadow
maps that offer an alternative, dynamic and highly detailed account
of the spatial dynamics of even very remote places where accurate
mapping has previously been limited or nonexistent. They also pre-
sent new challenges of interpretation which demand new skills and
levels of contextual understanding from researchers, since different
technologies make people visible in different ways. For example,
mobile phone location data displays people’s movement and activi-
ties, but not their interactions with their physical surroundings
(Taylor, 2015a), whilst satellite data shows people’s physical context,
but does not offer insights into what they may be doing. As
Nathaniel Raymond of the Harvard Humanitarian Initiative has said,
‘[there is a] lack of specific humanitarian pedagogy for interpreting
this data’,
so that those in possession of detailed maps formed for
use in humanitarian or longer-term development work are effec-
tively flying blind in terms of how to deal with it.
The implications of the new shadow maps are far-reaching. Just
as military maps have been put to civilian use, so methods for gen-
erating spatial data may see function creep from civilian to military
uses. Similarly, these maps-as-byproducts of other processes are
likely to find uses beyond their original purposes. For example,
the Harvard Humanitarian Initiative’s Satellite Sentinel Project
(SSP) uses satellite data to map and monitor human rights abuses
on Sudan’s southern border. Nathaniel Raymond, leader of the ini-
tiative, relates:
‘What we realised with Satellite Sentinel was that we were
changing the battle space. We had gone from this idea that we
were passive observers – our monitoring hat ... [to] changing
the tempo at which the emergency happens and the very nature
and outcome of the emergency itself. It is doing that by providing
unique situational awareness otherwise not available to armed
actors in real time. There is intelligence-grade situational aware-
ness in the case of sensors: what can be used to document a
human rights abuse can also be used to target an artillery strike.’
[Nathaniel Raymond, interview 25.2.2015]
Spatial data projects may be dedicated to providing security,
development planning information or reaching a new customer
base, yet they increasingly provide as byproducts data that can
map some of the world’s remotest places. A humanitarian, peace-
keeping, business or ‘development’ drone in a remote and virtually
unmapped area is inevitably collecting data not only on its targets
but on everyone. Spatial data derived from drones, mobile phone
records, social media and other emerging sources offer real-time,
dynamic information on where people are located and what they
are doing socially, economically and politically. Such datasets
may have particular power where they present an alternative map-
ping of low-income and remote communities, in ways which have
already been recognised as highly sensitive in HICs (ACLU, 2011).
As Raymond explains, in the case of digital humanitarianism:
‘You have data that is publicly available that by itself is not
weaponised to be harmful. But then digital humanitarians come
in and start creating cross-cutting correlating datasets that
allow seemingly static data [...] to turn into a bomb.’
[Interview, 25.2.2015]
The new shadow-maps and country data doubles are the results
of a structural change in the way digital data is produced and used.
Corporations are now embedded to an unprecedented extent in the
production of data that can be used to track, monitor, map and
analyse activities in LMICs, and their data trickles upward towards
more powerful, technologically adept collaborators such as the UN
(e.g. Global Pulse), researchers (such as the Harvard Humanitarian
Initiative), development donors and, inevitably, national intelli-
gence services. This complex structure of information gatherers
and users gives rise to a diverse set of visibilities which underpin
more distributed governance structures. Multiple possible futures
are created when others access these data, both good and bad –
but in every case power accrues to wealthier international actors,
who have the access to the data and the capacity to analyse it, at
the expense of less well-resourced local ones who do not. Thus
the deep embedding of corporations in processes of creating coun-
try data doubles creates a default where those doubles are not
accessible to states, and where mapping, counting and sorting
become distributed rather than centralised functions.
4. Conclusion
We have outlined two trends which demonstrate an institu-
tional and political power shift with regard to digital data in
LMICs. First, the way in which practices of informational capitalism
are shaping development interventions via public–private
Nathaniel Raymond, Harvard Humanitarian Initiative, presentation at Humani-
tarian Innovation Conference, 19–20 July 2014, Keble College, Oxford.
L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237 235
partnerships. Multinational technology firms are involved in inter-
ventions to map, sort and categorise, often also connecting people
to new technologies for the first time, and in return capture data on
new technology users and, through that data, new consumers and
markets. The second trend is for the visibilities created by these
corporate interventions to constitute new population-level data-
bases and maps. We characterise these new data infrastructures
as shadow maps and state data doubles because they operate in
parallel to state data collection infrastructures, and give rise to dis-
tributed forms of governance and powers to intervene.
This power shift from state to corporate data collectors and pro-
cessors suggests that there is a need to move beyond Scott’s influ-
ential 1998 framing of population-level data collection as
providing legibility and governability. In the cases we have
explored, the visibility provided by big data offers a ‘god’s eye
view’ (Pentland, 2011) in terms of potential scale and detail –
but if the new data create gods, they are gods with limited under-
standing. Unlike survey data, the new visibilities reflect those using
particular digital technologies or services, and without analogue
sources such as censuses and qualitative research to provide a
baseline or contextual information, there is no way to pin down
their bias and understand what they represent (González-Bailón
et al., 2012). Despite offering an unprecedented level of detail on
individual activities, movements and behaviour, big data is there-
fore also very fuzzy as a basis for intervention. Furthermore, the
new visibilities are different from legibility in that the data are col-
lected by unaccountable third parties rather than states, and there-
fore do not offer the potential for increasing accountability or
What, then, are the risks of such visibilities in a context of
development? First, a lack of accountability may give rise to a cul-
ture of experimentation amongst technology giants (as has already
been seen in the mass experiments conducted on users by various
online services such as Facebook and OKCupid). Greenleaf (2013:
11) shows that the overwhelming majority of states without pri-
vacy or data protection laws are low- or lower-middle-income
countries. For example, in Sub-Saharan Africa in 2013, only 8 states
out of 55 had data protection laws (ibid). Although many countries
are adopting versions of the EU or US data protection standards,
these do not so far restrict corporations from sharing data for
research or humanitarian reasons. If big data analytics in the name
of Development remain effectively unregulated, human-subject
experiments framed as development are more than a possibility.
Reid-Henry (2011) has warned of the use of ‘the global borderlands
as an appropriate site for experimentation in the government of
peoples’: in this paper we have charted one route that this type
of experimentation may take.
Next, if visibility distributes the power to govern and dimin-
ishes accountability, it also creates the power to represent data
subjects. Individuals cannot control how they are represented,
however, where their data flows through private sector systems
and is then sold on or given out pro bono to third parties.
Big-data representation occurs on an aggregate scale and the sub-
ject has no control over – or often knowledge of – their visibility.
Similar problems apply on the state level: representing countries
using big data has implications for state sovereignty, given that
big data can make visible economic and social dynamics which
are relevant to state concerns but which are not collected using
state mechanisms or necessarily shared with states. In PPPs or
under conditions where corporations act as development donors,
states are not – or are only partly – in charge of collecting and dis-
seminating these representations of themselves.
A further risk posed by the new data visibilities is that of what
Morozov (2013) has termed ‘solutionism’: the application of engi-
neering solutions to problems that are long-term and structural in
nature. The problems the new data are being applied to are old and
deeply embedded, and it is fair to ask whether corporations can
offer anything new in terms of the far-reaching structural change
that is necessary to ‘solve’ problems such as disease, poverty and
inadequate sanitation – or whether the language of ‘solutions’ is
even relevant in this domain. Responsible data science is being
conducted using commercial data (e.g. Bengtsson et al., 2011,
which used mobile data to map cholera transmission in Haiti in
the aftermath of the 2011 earthquake), but despite the positive
potential of the new shadow-maps and data doubles, so far find-
ings have yet to connect with action (Talbot, 2014). If the author-
ities with the resources and capacity to use better data do not
act on it, the data simply constitutes, as Nathaniel Raymond of
the Harvard Humanitarian Initiative has said, ‘a post-mortem’ on
problems in LMICs.
Last, perhaps the most serious risk is that informational capital-
ism becomes the new norm for the field of international develop-
ment. States and traditional development donors have until now
‘seen like states’, using legibility through conventional statistics
as a basis for intervention. Yet today, if people are most visible to
institutions with the most data and analytical power, LMIC citizens
are now seen most clearly by multinational technology corpora-
tions. Meanwhile, the possession of unprecedentedly extensive
data is empowering the private sector as primary actors, rather
than contractors, in development planning and interventions. The
shift from legibility to visibility – and from public-sector reading
to privatised seeing – requires both new theory and a new ethical
approach. In theoretical terms, development interventions may
increasingly become a byproduct of informational capitalism,
rather than a way for states to achieve domestic or foreign policy
aims, and would therefore need to be assessed according to differ-
ent criteria and with a different institutional focus. Ethically, if
LMICs are becoming the new laboratories of data science, there is
a case to be made for closely monitoring the boundary between
care and control; for determining responsible bodies to regulate
and monitor the data science being conducted on LMIC subjects;
and particularly for reevaluating the nature and priorities of pri-
vacy in an LMIC context, where – as we have shown – accountabil-
ity for data use and intervention may be minimal.
This work was supported by the European Union’s Marie Curie
postdoctoral program, Grant #624583.
ACLU, 2011. Protecting Privacy from Aerial Surveillance: Recommendations for
Government Use of Drone Aircraft. Online publication, accessed 15.08.14 at
Amoore, L., 2011. Data derivatives. On the emergence of a security risk calculus for
our times. Theory Cult. Soc. 28 (6), 24–43.
Bauman, Z., Bigo, D., Esteves, P., Guild, E., Jabri, V., Lyon, D., Walker, R.B.J., 2014.
After snowden: rethinking the impact of surveillance. Int. Political Sociol. 8 (2),
BBC News, 2012. Matternet: swapping roads for flying drones. Accessed 15.08.14
at <
Bengtsson, L., Lu, X., Thorson, A., Garfield, R., von Schreeb, J., 2011. Improved
response to disasters and outbreaks by tracking population movements with
mobile phone network data: a post-earthquake geospatial study in Haiti. PLoS
Med. 8 (8), e1001083.
Berlingerio, M., Calabrese, F., Di Lorenzo, G., Nair, R., Pinelli, F., Sbodio, M.L., 2013.
AllAboard: a system for exploring urban mobility and optimizing public
transport using cellphone data. Mobile Phone Data for Development: analysis
of mobile phone datasets for the development of Ivory Coast. Netmob
conference, 1–3 May 2013, MIT, Boston (also published in Machine Learning
and Knowledge Discovery in Databases (pp. 663–666). Springer, Berlin
Bowker, G., Star, S., 2000. Sorting Things Out. Classification and Its Consequences.
The MIT Press, Cambridge, Massachusetts.
Burns, R., 2014. Rethinking big data in digital humanitarianism: practices,
epistemologies, and social relations. GeoJournal, 1–14.
236 L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237
Business Standard News, 2013. ‘UIDAI has issued 400 mn Aadhaar numbers till
date’. Accessed 06.10.14 at <
Castells, M., 1996. The Rise of the Network Society. The Information Age: Economy,
Society and Culture, vol. I. Blackwell, Oxford, UK.
Center for Global Development, 2014. Delivering on the Data Revolution in Sub-
Saharan Africa. Accessed online 17.9.2014 at <
Cherlet, J., 2014. Epistemic and technological determinism in development aid. Sci.
Technol. Human Values 39 (6), 773–794.
Churchill, E., 2014. Philippines: drones and spreadsheets to the rescue. <http://
Cohen, J., 2013. What privacy is for. Harv. L. Rev. 126, 1904.
Crampton, J.W., Roberts, S.M., Poorthuis, A., 2014. The new political economy of
geographical intelligence. Ann. Assoc. Am. Geogr. 104 (1), 196–214.
Crowe, A., 2013. United Nations’ drones: A sign of what’s to come? Accessed
15.08.14 at <
Degli Esposito, S., 2014. When big data meets dataveillance: the hidden side of
analytics. Surveill. Soc. 12 (2), 209–225.
Dreze, J., 2010. Unique facility, or recipe for trouble? The Hindu, November 25,
2010. <>
(accessed 16.09.14).
EFF, 2015. Is Not Neutral, Not Secure, and Not the Internet.
Electronic Frontier Foundation, May 18, 2015. Accessed 15.6.2015 at <https:// -not-neutral-not-secure-and-not-
Global Pulse, 2012. Big Data for Development: Opportunities & Challenges.
White Paper. Accessed 22.08.14 at <
Global Pulse, 2013. United Nations Global Pulse Information Sheet. <www.>.
Global Pulse, 2014. Mining Indonesian tweets to understand food price crises.
Methods paper. Accessed 07.08.14 at <
González-Bailón, S., Wang, N., Rivero, A., Borge-Holthoefer, J., Moreno, Y., 2012.
Assessing the bias in communication networks sampled from twitter. Available
at SSRN 2185134.
Graham, M., 2009. Neogeography and the palimpsests of place: web 2.0 and the
construction of a virtual earth. Tijdschrift voor Economische en Sociale
Geografie 101 (4), 422–436.
Green, D., 2014. Big data and development: upsides, downsides and a lot of
questions. Accessed 04.08.14 at: <
Greenleaf, G., 2013. Sheherezade and the 101 data privacy laws: origins,
significance and global trajectories. J. Law Inf. Sci. 23 (1).
Grumbach, S., Frénot, S., 2013. Les données, puissance du futur. Le Monde. Accessed
5.7.2016 at <
GSMA (2013b), 2013: the year in research. Accessed 22.08.14 at <https://
Haggerty, K.D., Ericson, R.V., 2000. The surveillant assemblage. Br. J. Sociol. 51 (4),
Harley, J., 1989. Deconstructing the map. Cartographica 26 (2), 1–20.
Hildebrandt, M., 2013. Slaves to Big Data. Or Are We? Idp. Revista De Internet,
Derecho y Política 16.
Hosein, G., Nyst, C., 2013. Aiding Surveillance: An Exploration of How Development
and Humanitarian Aid Initiatives are Enabling Surveillance in Developing
Countries. Available at SSRN 2326229.
IBM, 2014. Project Lucy: How IBM Research is engaging in Africa. White paper, IBM,
New York. Accessed 25.08.14 at <
IndiaToday, 2013. Aadhaar scheme: How safe and secure is your identity? Accessed
12.08.14 at <
secure-is-your-identity/1/313616.html>., 2014. Announcing the Connectivity Lab at Facebook. Accessed 07.08.14
at < lab-at-
Jayaram, M., 2014. India’s Big Brother Project. Boston Review, May 19th 2014.
Accessed online 06.10.14 at <
Jerven, M., 2013. Poor Numbers: How We Are Misled by African Development
Statistics and What to Do about It. Cornell University Press, Ithaca.
Kirkpatrick, R., 2011. Data philanthropy: public and private sector data sharing for
global resilience <
Kitchin, R., 2014. The real-time city? Big data and smart urbanism. GeoJournal 79,
Krishnan, R., 2012. Aadhaar: A National Innovation Platform? Accessed 06.10.14 at
LiveMint, 2013. Nasscom, UIDAI to push Aadhaar app development. Accessed online
25.08.14 at <
Livingstone, S., Walter-Drop, G., 2014. Bits and Atoms: Information and
Communication Technology in Areas of Limited Statehood. OUP, Oxford.
Lyon, D., 2007. Surveillance Studies. An Overview. Polity Press, Cambridge.
Lyon, D., 2014. Surveillance, Snowden, and Big Data: Capacities, consequences,
critique. Big Data Soc. July–September, 1–13.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., Hung Byers, A.,
2011. Big Data: The Next Frontier for Innovation, Competition and Productivity.
McKinsey Global Institute, Washington DC.
Mao, H., Shuai, X., Ahn, Y.Y., Bollen, J., 2013. Mobile communications reveal the
regional economy in Côte d’ Ivoire. In: Proceedings of the 3rd Conference on the
Analysis of Mobile Phone Datasets (NetMob).
Mayer-Schönberger, V., Cukier, K., 2013. Big data: a revolution that will transform
how we live, work, and think. Houghton Mifflin Harcourt.
Miraftab, F., 2004. Public–private partnerships the Trojan horse of neoliberal
development? J. Plan. Educ. Res. 24 (1), 89–101.
Morozov, E., 2013. To save everything, click here: Technology, solutionism, and the
urge to fix problems that don’t exist. Penguin, UK.
MTN, 2010. MTN launches Facebook
ZERO, a FREE way to access Facebook
your Mobile. Accessed 07.08.14 at <
Netmob, 2013. In: Proceedings of the 3rd Conference on the Analysis of Mobile
Phone Datasets, May 1–3, MIT, Boston.
Nilekani, N., 2013. Technology to Leapfrog Development: The Aadhaar Experience.
Richard H. Sabot lecture, Center for Global Development, April 22 2013. <http://
technology-leapfrog-development.pdf> (accessed 16.09.14).
Orange, 2012. D4D project: <>.
Pentland, A., 2011. Society’s nervous system: building effective government, energy,
and public health systems. Pervasive Mobile Comput. 7 (6), 643–659.
Polzer, T., Hammond, L., 2008. Invisible displacement. J. Refug. Stud. 21 (4), 417–431.
Ramanathan, U., 2013. Who Owns The UID Database? Medianama, <http://www.>
(accessed 16.09.14).
Reid-Henry, S., 2011. Spaces of security and development an alternative mapping of
the security–development nexus. Secur. Dialogue 42 (1), 97–104.
Ruppert, E., 2012. The governmental topologies of database devices. Theory Cult.
Soc. 29 (4–5), 116–136.
Schneier, B., 2015. Data and Goliath. The Hidden Battles to Collect Your Data and
Control Your World. W.W. Norton and Company, New York.
Scott, J.C., 1998. Seeing Like a State: How Certain Schemes to Improve the Human
Condition Have Failed. Yale University Press.
Shearmur, R., 2015. Dazzled by data: big data, the census and urban geography.
Urban geography.
Talbot, D., 2014. Cell-Phone Data Might Help Predict Ebola’s Spread. MIT
Technology Review. Accessed 17.09.14 at <http://www.technologyreview.
com/news/530296 /cell-phone-data-mig ht-help-predict-ebolas-spread/? utm_
content=buffer05d6d &utm_ medium=social&utm_sou rce=t witte
Taylor, L., 2015a. No place to hide? The ethics and analytics of tracking mobility
using mobile phone data. Accessed 30.06.15 at <
Taylor, L., 2015b. A god’s eye view? Big data, ground truth and the D4D challenge.
Accessed 30.06.15 at <
Taylor, P.J., Johnston, R.J., 1995. Geographic information systems and geography. In:
Pickles, J. (Ed.), Ground Truth.
Taylor, L., Schroeder, R., 2014. Is bigger better? The emergence of big data as a tool
for international development policy. GeoJournal, 1–16.
Thomas, O.D., 2014. Foucaultian dispositifs as methodology: the case of anonymous
exclusions by unique identification in India. Int. Political Sociol. 8 (2), 164–181., 2014. About UIDAI. Accessed 12.08.14 at <
Volkmann, W., 2014. Small Drones for Progress in International Development
Projects. Post on LinkedIn, April 24, 2014. Accessed 15.08.14 at <https://www.
World Economic Forum, 2012. Big Data, Big Impact: New Possibilities for
International Development. Accessed 22.08.14 at <
Yadav, V., 2013. Unique Identification Project for 1.2 billion People in India. SAI
Working Paper, Harvard University.
Zelazny, F., 2012. The Evolution of India’s UID Program. Lessons Learned and
Implications for Other Developing Countries. Center for Global Development
Policy Paper.
L. Taylor, D. Broeders / Geoforum 64 (2015) 229–237 237
... What few commentators on big data forecasted was the extent to which big data would represent a private sector revolution (Taylor & Broeders, 2015). The step change in volume, immediacy and power constituted by the new sources of large-scale data did not stem from bureaucratic or academic innovation, but from changes in the commercial world driven by new devices and massive investments in software, hardware and infrastructure. ...
Full-text available
Big data has increased attention to Computational Social Science (CSS) on the part of policymakers because it has the power to make populations, activities and behaviour visible in ways that were not previously possible. This kind of analysis, however, often has unforeseen implications for those who are the subjects of the research. This chapter asks what a social justice perspective can tell us about the potential, and the risks, of this kind of analysis when it is oriented towards informing policy. Who benefits, and how, when computational methods and new data sources are used to conduct policy-relevant analysis? Should CSS sidestep, through its novelty and its identification with computational and statistical methodologies, sidestep ethical review and the assessments of power asymmetries and methodological justification that are common in social science research? If not, how should these be applied to CSS research, and what kind of assessment is appropriate? The analysis offers two main conclusions: first, that the field of CSS has evolved without an accompanying evolution of debates on ethics and justice and that these debates are long overdue. Second, that CSS is privileged as policy-relevant research precisely because of many of the features which bring up concerns about justice—large-scale datasets, remote data gathering, purely quantitative methods and an orientation towards policy questions rather than the needs of the research subjects.
... This shift deserves increased scholarly attention, because the academic fields of computer science and engineering, where most new technology emerges from, and the private sector companies that utilize these technologies, do not necessarily share the same concerns or objectives as conventional peacebuilding practitioners and experts. In the adjacent field of international development, "datafication" has led to an increasing involvement of the private sector in knowledge generation, reducing the relevance of publicly owned data and enabling new forms of intervention based on fine-grained representations of target populations ( Taylor and Broeders 2015 ). This trend is also visible in international efforts to counter hate speech, where lexica developed through local initiatives may be commonly utilized to develop artificial intelligence (AI) and especially natural language processing (NLP) algorithms that support automated hate speech detection on social media platforms ( Poletto et al. 2021 ). ...
Full-text available
Existing research on digital technologies in peacebuilding exhibits both tech-solutionist and tech-problematizing traits that tend to understate their embeddedness in society and politics. We argue that the study of digital peacebuilding should instead reflexively engage with the coproduction of the technical and the social in both academia and practice. This requires asking how assumptions about technology are related to assumptions about the conflict and peacebuilding context on which these technologies are brought to bear, and with what consequences. Therefore, we propose a methodological framework that brings to the fore how technologies for peacebuilding and peacebuilding with technology are coproduced. First, we focus on the interrelated claims about peacebuilding and technology, and the coproduction of peacebuilding problems and technological solutions. Second, we inquire into the characteristics of the digital peacebuilding agendas built on these claims, including the dynamics of disruptive change and datafication that these agendas bring. Third, we consider the sticky effects of digital approaches, in terms of a politicization or depoliticization of peacebuilding efforts, and ask what kind of peace this may produce.
... Ownership of data in these environments is also a difficulty with power shifting from traditional development actors to corporations and public-private partnerships. While Kwet (2019) characterised United States (US) dominance of the digital ecosystem in many African countries as an insidious form of "digital colonialism", Taylor and Broeders (2015) observed how datafication is affecting power dynamics in the Global South. Concerned at the growth in data collection and processing by large corporations in these countries, they identify the risk of development interventions becoming a by-product of "informational capitalism". ...
Collecting relevant and appropriate data when conducting research in the Global South and other resource-constrained environments can be challenging. This is particularly true when the researcher originates from a different country and has different social, cultural, and political beliefs to the research participants. Data collection in such contexts can be challenging for a variety of ethical, philosophical, theoretical, and methodological reasons which may arise resulting from the use of research approaches designed in the Global North. There are also many practical and infrastructural issues which may be relevant in resource-constrained countries. This paper provides examples of challenges encountered by the authors in a variety of ongoing research projects in the Global South. We propose approaches to address the identified challenges, and we conclude by calling for additional work on this important topic.
... Research on the institutional work of digital platforms suggests that they enable an aggregation by the platform and its owners of market institutional functions and power that were previously distributed and dissipated . Processes of datafication associated with many current digital systems tend to be inherently unequal because users necessarily give over their data to the system owners, but there is no inherent basis for the reverse process of making the owners' data transparent (Taylor & Broeders, 2015). Digital has alsovia machine learning and algorithmsmade systemic processes such as decision-making or distribution of value more opaque, but particularly to system users rather than owners (Burrell, 2016). ...
Full-text available
Digital systems are significantly associated with inequality in the global South. That association has traditionally been understood in terms of the digital divide or related terminologies whose core conceptualization is the exclusion of some groups from the benefits of digital systems. However, with the growing breadth and depth of digital engagement in the global South, an exclusion worldview is no longer sufficient. What is also needed is an understanding of how inequalities are created for some groups that are included in digital systems. This paper creates such an understanding, drawing from ideas in the development studies literature on chronic poverty to inductively build a model of a new concept: ‘adverse digital incorporation’, meaning inclusion in a digital system that enables a more-advantaged group to extract disproportionate value from the work or resources of another, less-advantaged group. This new model will enable those involved with digital development to understand why, how and for whom inequality can emerge from the growing use of digital systems in the global South. It creates a systematic framework incorporating the processes, the drivers, and the causes of adverse digital incorporation that will provide detailed new insights. The paper concludes with implications for both digital development researchers and practitioners that derive from the model and its exposure to the broader components of power that shape the inclusionary connection between digital and inequality.
This paper examines the knowledge politics and cultures that shape data relations between international aid organisations and Global South public institutions, taking African libraries as an example. International organisations increasingly rely on data from the Global South, purportedly as a resource for development, which has raised valid concerns about the emergence of new practices of data colonialism. One proposed solution is to expand the capacity of Global South institutions to control their own data processes, so they can likewise control the politico-economic relationships that draw on their data. A pan-African library organisation representing 34 countries is exploring this possibility though a multiyear research project to increase library capacity to use data to partner with development aid organisations. However, this work revealed that data colonialism precedes practices of value extraction. In focus groups, a survey of library systems and interviews with aid organisations, aspects of the data cycle are epistemically framed by aid organisations to undercut Global South control, and subtle neocolonial mechanisms encourage libraries to shape their own data cultures according to desires of aid organisations. This underscores the need to expand data neocolonialism as a frame for confronting epistemic injustice by highlighting Western rationalities embedded in data relations.
Research in HCI4D has continuously advanced a narrative of ‘lacks’ and ‘gaps’ of the African perspective in technoscience. In response to such misguided assumptions, this paper attempts to reformulate the common and perhaps unfortunate thinking about African practices of design in HCI4D – i.e., largely as a function of African societal predicaments and Western technocratic resolutions. Through critical reflection on a range of issues associated with post-colonialism and post-development, I examine the possibilities that various historical tropes might offer to the reinvention of the African perspective on innovation. This leads to the consideration of how engaging in critical discussions about the future dimensions of African HCI can allow for grappling with the effect of the coloniality of being, power and knowledge. Developing on the ideas of futuring as a way of dealing with the complexities of the present – in this case the coloniality of the imagination - the paper ends by discussing three tactical propositions for ‘remembering’ future identities of African innovation where the values of autonomy are known and acted upon.
This paper asks whether datafication practices are dehumanising international development and if a human-centred and participatory datafication is possible. The paper uses Habermas’ theory of the different ‘knowledge interests’ that constitute different forms of social action. Three kinds of datafication projects are explored: humanitarian AI, digital-ID and community mapping. The authors argue that data-science and participatory practices are forms of social action that are shaped by different knowledge-interests. It is argued that the technical knowledge interests shaping datafication projects conflict with high-level policy commitments to participatory development. Ethical Principles of AI are assessed as a route to more human-centred practices of datafication for development. The authors argue that avoiding tokenistic forms of participation will require the incorporation of practical and emancipatory knowledge interests and the use of new monitoring and evaluation tools to trace the achieved levels of participation of different actors at each stage of the project cycle.
The Global Challenges Research Fund (GCRF) provided a mechanism for academia to undertake projects relevant to the Sustainable Development Goals but there have been limited opportunities to critically interrogate such projects. In this paper we will use the Technology Implementation Model for Energy to deconstruct the purpose, assumptions and expectations, engagement strategies, and reflective processes of four GCRF projects in order to better understand relationships between researchers and those being researched. Thus, the aim of this paper is to explore and understand the lived experiences of four inter-disciplinary GCRF Primary Investigators implementing poverty alleviating technologies in a range of sectors to generate recommendations that can be applied to wider academic communities engaging with vulnerable populations. Our key findings show that despite the integration of Theory of Change models and the Responsible Research and Innovation (RRI) framework in GCRF-funded projects, project aims continue to be driven by researchers rather than reflecting end-user needs. Whilst some projects looked to generate feelings of ownership, adequate engagement strategies and reflective learning practices, these processes are often not formally embedded in project activities resulting in a decoupling of researcher expectations and end-user assumptions – ultimately derailing project outcomes. Our recommendations for academics operating within the International Development space are to 1) Talk early, often and transparently, 2) Keep Thinking – who benefits?, 3) Be reflective, responsive, and open to change and, 4) Use a systematic approach to facilitate this process.
This chapter considers the social conditions in which large-scale biometric systems have been deployed in emerging economies across three cases: Jamaica, Afghanistan, and Kenya. Its contributions to the study of biometrics and forensics are both empirical and theoretical. The empirical contribution rests on the attention to comparatively under-researched geographies and political processes of technology-driven social transformation in the Caribbean, central Asia, and east Africa. The theoretical contribution rests on the elaboration of sociopolitical factors that have hampered the effective uptake of these technologies as well as engagement in dialogue with the body of literature on development-driven technological interventions into the governance of emerging economies. By undertaking a critical review of these contemporary cases, the chapter presents the state of the art in both theory and implementation while illustrating the necessities of popular legitimacy, equitable access, universal registration, and clearly elaborated data protection regimes in biometric rollouts.KeywordsBiometricsTechnocolonialismJamaicaAfghanistanKenya
Traditionally, encryption keys are stored in a cryptosystem, risking theft or loss. Biometric encryption is a possible solution to avoid the need to store keys securely. The basic idea is that a cryptographic key can be generated from or bond with the biometric data of a user whenever a key is needed. This chapter aims to review and evaluate the state-of-the-art biometric encryption techniques. Different techniques to acquire, process, and extract data from iris biometric samples are evaluated and compared. Three template-free techniques are designed and implemented to test the performances of the system in terms of false acceptance rate (FAR) and false rejection rate (FRR). The results show that it is possible to generate an ECC (error correcting code) key pair and identify a person with a 3.7% FAR and a 21% FRR, values that can be further improved by optimizing the initial processing of the iris.
Full-text available
Spatial technologies and the organizations around them, such as the Standby Task Force and Ushahidi, are increasingly changing the ways crises and emergencies are addressed. Within digital human-itarianism, Big Data has featured strongly in recent efforts to improve digital humanitarian work. This shift toward social media and other Big Data sources has entailed unexamined assumptions about technological progress, social change, and the kinds of knowledge captured by data. These assumptions stand in tension with critical geographic scholarship, and in particular critical GIS research. In this paper I borrow from critical research on technologies to engage three important new facets of Big Data emerging from an interrogation of digital humanitarianism. I argue first that within digital humanitarianism, Big Data should be understood as a new set of practices, in addition to its usual conception as data and analytics technologies. Second, I argue that Big Data constitutes a distinct epistemology that obscures many forms of knowledge in crises and emergencies and produces a limited understanding of how a crisis is unfolding. Third, I argue that Big Data is constitutive of a social relation in which both the formal humanitarian sector and ''victims'' of crises are in need of the services and labor that can be provided by digital humanitarians.
This paper examines the ethical and methodological problems with tracking human mobility using data from mobile phones, focusing on research involving low- and middle-income countries. Such datasets are becoming accessible to an increasingly broad community of researchers and data scientists, with a variety of analytical and policy uses proposed. This paper provides an overview of the state of the art in this area of research, then sets out a new analytical framework for such data sources that focuses on three pressing issues: first, interpretation and disciplinary bias; second, the potential risks to data subjects in low- and middle-income countries and possible ethical responses; and third, the likelihood of ‘function creep’ from benign to less benign uses. Using the case study of a data science challenge involving West African mobile phone data, I argue that human mobility is becoming legible in new, more detailed ways, and that this carries with it the dual risk of rendering certain groups invisible and of misinterpreting what is visible. Thus, this emerging ability to track movement in real time offers both the possibility of improved responses to conflict and forced migration, but also unprecedented power to surveil and control unwanted population movement.
The use of digital communication technologies, and of mobile phones in particular, has seen an exponential rise in low- and middle-income countries over the last decade. These data, emitted as a byproduct of technologies such as mobile phone location information and calling metadata, have the potential to fill some of the problematic gaps in data resources available to country policymakers and international development organisations. Using three examples of current big data initiatives in the international development field, we examine the implications of these new types of data for development policy and planning: their advantages and drawbacks, emerging practices relating to their use, and how they potentially influence ideas and policies of development. We also assess the politics of these new types of digital data, which are often collected and processed by corporations or by researchers in industrialised countries. Our analysis indicates that these new data sources already represent an important complement to country-level statistics, but that there are currently important challenges which will need to addressed if the promises of big data in development are to be fulfilled.
Since the turn of the millennium, the major development agencies have been promoting "knowledge for development,'' "ICT for development,'' or the "knowledge economy'' as new paradigms to prompt development in less-developed countries. These paradigms display an unconditional trust in the power of Western technology and scientific knowledge to trigger development-they taste of epistemic and technological determinism. This article probes, by means of a genealogy, how and when development cooperation began adhering to epistemic and technological determinism, and which forms this adhesion has taken over time. The genealogy shows, first, that knowledge and technology have always been integrally part of the very "development'' idea since this idea was shaped during enlightenment. Second, while the genealogy reveals that epistemic and technological determinism were embedded in the development idea from the very beginning, it also illustrates that the determinism has always been challenged by critical voices.