Data Management and Use: Governance in the 21st Century

I was on the working group for this report that was published in 2017 by the British Academy and Royal Society.

Responding to the challenges of data in the 21st Century, the report concluded:

“We think two responses are required. Firstly, a set of high-level principles is needed to visibly shape all forms of data governance and ensure trustworthiness and trust in the management and use of data as a whole. The promotion of human flourishing is the overarching principle that should guide the development of systems of data governance. The four principles that follow provide practical support for this overarching principle: all systems of data governance across the varied ways data is managed and used should,

  • protect individual and collective rights and interests
  • ensure that trade-offs affected by data management and data use are made transparently, accountably and inclusively
  • seek out good practices and learn from success and failure
  • enhance existing democratic governance.”

Ethnography for a Data-Saturated World

This book that I edited with Dawn Nafus, published in October 2018, tackles many of the issues I’m exploring in this research project

The blurb reads:

“This edited collection aims to reimagine and extend ethnography for a data-saturated world. The book brings together leading scholars in the social sciences who have been interrogating and collaborating with data scientists working in a range of different settings. The book explores how a repurposed form of ethnography might illuminate the kinds of knowledge that are being produced by data science. It also describes how collaborations between ethnographers and data scientists might lead to new forms of social analysis”


1. Introduction: ethnography for a data-saturated world – Hannah Knox and Dawn Nafus
Part I: Ethnographies of data science
2. Data scientists: a new faction of the transnational field of statistics – Francisca Grommé, Evelyn Ruppert and Baki Cakici
3. Becoming a real data scientist: expertise, flexibility and lifelong learning – Ian Lowrie
4. Engineering ethnography – Kaiton Williams
Part II: Knowing data
5. ‘If everything is information’: archives and collecting on the frontiers of data-driven science – Antonia Walford
6. Baseless data? Modelling, ethnography and the challenge of the Anthropocene – Hannah Knox
7. Operative ethnographies and large numbers – Adrian Mackenzie
Part III: Experiments in/of data and ethnography
8. Transversal collaboration: an ethnography in/of computational social science – Mette My Madsen, Anders Blok and Morten Axel Pedersen
9. The data walkshop and radical bottom up data knowledge – Alison Powell
10. Working ethnographically with sensor data – Dawn Nafus
11. The other 90%: thinking with data science, creating data studies – Joseph Dumit interviewed by Dawn Nafus

And Sarah Pink has kindly provided a great endorsement for the book:

‘Ethnography for a Data Saturated World is a must-read for researchers, students and professionals outside academia wishing to understand what digital data means for our contemporary world. It brings our attention to a burgeoning field of research and practice which unites ethnography and data science on a number of levels. This book takes us into the world of digital data in a mode and depth that only the particular sensibilities of ethnographic research can offer. Its editors and authors collectively provide a new and global vision through ethnographic studies of how the worlds of data scientists are constituted, the ways of knowing and forms of expertise that digital data analysis involves, and the methodological challenges and achievements of work that has created new modes of collaboration between ethnography and digital data analysis. Ethnography for a Data Saturated World is at once a substantive, theoretical and methodological book. It is brimming with significant ethnographic insights and findings about the worlds it examines, it offers an array of different and disciplinary specific modes of thinking theoretically about digital data from anthropology and sociology, and it interrogates the modes of knowing that are implicated in both digital data collection and analysis and in ethnographic practice, as well as the possible connections between them.’
Sarah Pink, Professor of Design and Media Ethnography, RMIT University


What can anthropology tell us about data in the 21st century?

This piece was published in January 2018 on the Observatory for a Connected Society app:

What can anthropology tell us about data in the 21st century?

17 January 2018
Data is no longer just the outcome of scientific research or administrative functions of government but is now created as a bi-product of every person’s interactions with the internet, infrastructures, institutions, news media, supermarkets, banks, the built environment and so on. Building on the work of anthropologists who have been trying to make sense of data and its social implications, Dr Hannah Knox makes the case for the crucial role that anthropology can play in wading through this data saturated landscape.
It is a truth commonly expressed that we live in a world saturated by digital data. Data is no longer just the outcome of scientific research or administrative functions of government but is now created as a bi-product of every person’s interactions with the internet, transport infrastructures, institutions, news media, supermarkets, banks and the built environment. Confronted with this ever-increasing mass of digital data there is both excitement and consternation about how this data should be analysed and what the implications of its use will be for the future of work, knowledge and social relations.My academic focus is on data and its social implications and I see a crucial role for anthropologists in helping others make sense of this data-saturated landscape. Anthropologists are uniquely equipped to explore the promises and expectations of data and to understand their effects. Far too often, commentaries on the social promise or cultural dangers of data are dominated by technicist accounts that fail to appreciate the way in which digital data, even in its most-posthuman manifestations (e.g. general artificial intelligence or advanced robotics) remains a deeply human endeavour.

When we approach data from the perspective of technical systems we are confronted with what looks like enormous complexity – algorithms working invisibly in the background using things like Bayesian techniques for determining probabilistic relations, gaussian prediction and github repositories of code running to hundreds of pages to produce links and insights that encourage us to buy, click and skim (1). But what if we were to try to understand data from the position of the people that work with it and manage it? What other kinds of understandings of this data landscape would this elicit?

This is what anthropologists are beginning to do. Anthropologists are sometimes criticised for pointing out that things are simply ‘more complex’ than they seem at first sight. But in the case of digital data, I would argue that anthropology offers a way of re-describing data, through an attention to human practices and ideas, so as to make it less complex and more understandable for those who are not steeped in the technical languages of coding, mathematics or computer engineering.

Anthropologists are experts in translation. The classic image of the anthropologist is that of the intrepid explorer visiting far-off cultures to bring back tales of the exotic rendered comprehensible through social analysis. In fact, today you are as likely to find an anthropologist in a science laboratory, a government office, a protest march or a community allotment as you are to find them hanging out in a Papua New Guinean village. But whether doing research in far off places, or in social situations that seem closer to home, anthropologists are always cultural translators, turning the seemingly incomprehensible dimensions of the worlds they study into terms that other anthropologists, and hopefully others who are not anthropologists, can begin to understand.

Being an anthropologist invariably involves learning another language. To do our research we must become competent members of a community learning the terminologies, rituals and practices of the group of people we are studying. This process of gaining intimacy with a community, learning social and linguistic cues and becoming versed in the techniques that are often taken for granted can be an awkward experience, full of surprises and mistakes. Rather than papering over these failures and mistakes, or erasing the aspects of people’s activities that don’t fit preconceived ideas anthropologists use these experiences as a way of interrogating the difference between their own assumptions about the world, and the assumptions of those with whom they are doing research.

This allows anthropologists to unravel and unpack what is often taken for granted. When people say digital technologies will lead to the end of work, the perplexed anthropologist who might understand work as a social contract, will wonder how it can be that such a social contract could be imagined as disappearing. This might entail asking whether work for the person making such a statement is indeed a social contract, and it is that which is disappearing, or whether it is seen as something else which is disappearing, in which case the question becomes, what? With these kinds of questions, we begin to unravel what’s taken for granted in the everyday, allowing us to better understand just what it is that people fear or desire about digital technologies (skills, identity, continuity, community, safety, security?) and where those worries and hopes come from (sense of self, ethical stance, moral interpretation?).

The proliferation of digital data, and the challenges it poses, offers a fertile terrain for this kind of anthropological work. The current enthusiasm for blockchain, machine learning and predictive analytics, raises questions about precisely what it is that is driving this interest and what the effects of this enthusiasm are. In relation to blockchain, media studies scholar Lana Swartz (2) has shown, using precisely the analytic approach described, how interest in blockchain is sustained not only by the technical capacities of the distributed ledger but also operates as what she calls an ‘inventory of desire’. Focusing on what those working with blockchain actually say, rather than on an idealised version of what blockchain is supposed to do, Swartz shows the importance of liberal values of freedom, decentralisation, and privacy that underpin enthusiasm for blockchain, fuelling investment and development in the technology.

Similar analyses have been done on algorithms and the imaginaries that sustain them. Susanne Thompson and colleagues have recently gone so far as to suggest that algorithms might be usefully understood as ‘fetish’ objects. As they make clear, within anthropology, fetishes are understood not as ‘indices of false thinking’ but rather as ‘material objects that stabilise ongoing social relations because people invest them with [an] effect [of simultaneous belief and disbelief]’. Algorithms, they show, gain part of their power from their ability to both confirm people’s understanding of how the world should be, and to produce awe and wonder when they actually work.

This combination of belief in, and disbelief of, technology is perhaps most clearly evident in forms of data analysis that are oriented to the replication or improvement of human-like abilities. Machine learning, artificial intelligence and humanoid robots all entail a fetish-like form of engagement. Developers of intelligent machines draw explicitly on ideas about abilities that are derived from particular understandings of human being such as cognitive and rational capacities, haptic interaction, environmental awareness and logical deduction. When aspects of these qualities become replicated in computational machines, there is often a certain disbelief, awe and wonder expressed at the spectacle of a machine acting like a human.

The interest in finding ways of making machines act like humans have a long historical precedent which helps remind us that not everything about advanced data technologies in necessarily new. Current dreams of automation can be traced back to intricate automata made by Viennese clockmakers in the 17th century, through to Wolfgang Von Kempelen’s chess playing Mechanical Turk which turned out to be what Steven Shapin has insightfully described as device in which a human, pretended to be a machine that was pretending to be human (3). It is no coincidence that Amazon’s mechanical turk rests on a similar idea, whereby human beings stand in for machines that themselves are meant to replace humans.

For anthropologists, one of the most fascinating things about digital data is that the work to manage and manipulate it uncovers taken for granted ideas about human capacities. As well as being based on ideas about human capacities, computational machines are also shaping what we expect it means to be a human. Industrial manufacturing led to people becoming reimagined as units of productive labour. Now algorithms are leading to a rethinking of identity as a composition of experiences and preferences, and the gig-economy is making people rethink themselves not as units of labour but as marketable commodities in a shifting and unstable landscape of work. Whilst futurist predictions about digital data often worry about how computers are set to replace human beings, anthropological studies show how what makes us human is already being shaped by digital data and the machines that analyse it. The question is not whether robots or artificial intelligence will replace human beings but which kind of human they will replace, and with what implications.

In a report, published by the Royal Society and the British Academy last year, a call was made to ensure the governance of data puts human flourishing at its core. For anthropologists this is crucial, not just because it brings human beings into discussions about the benefits and dangers of digital technologies, but also because it allows us to talk about how digital technologies are framing what is valuable – about data, about machines but also, crucially, about what it is to be human in the world today.


  1. Adrian Mackenzie, 2017 ‘Infrastructures in Name Only: Identifying effects of Depth and Scale’ in Penny Harvey, Casper Bruun Jensen and Atsuro Morita (Eds) Infrastructures and Social Complexity
  2. Lana Swartz. 2017. “Blockchain Dreams: Imagining techno-economic alternatives after Bitcoin.” Another Economy is Possible, edited by Manuel Castells. Polity Press.
  3. Mechanical Marvels, Clockwork Dreams, BBC 4 documentary, Simon Schaffer, “Enlightened Automata” in The Sciences in Enlightened Europe, edited by William Clark, Jan Golinski, and Schaffer (Chicago University Press, 1999)


Conversations about Data

Data affects most aspects of contemporary life. Whilst data is not new, the proliferation of digital information that came first with computers, then the internet, and then ubiquitous computing, has transformed the question of what data is and how it affects people’s lives. Bits of digital information that are stored on servers and circulate as electrical pulses through wires are not just numbers or code but crucially, carry with them markers of identity, practice, preference, desire and meaning. Data matters so much then because it both confirms and potentially changes who we, as humans, are.

However most discussions of data gloss over this human side of data . Those who are enthralled by what data can do often fetishise it, giving it an agency all of its own as if code and numbers were able to float, circulate and aggregate freely without infrastructure, ownership, firewalls, or regulations.  When data fails to achieve its intended effects, these processes, technologies and ideas get blamed for polluting data and limiting an ideal of what it should be.

On the other hand, those who are committed to describing the richness of human experience have often seen data as a reduction, abstraction, and thin form of description compared to the depth that is possible through more literary forms of writing. Most anthropologists have avoided the study of data, but that is changing. A number of anthropologists with diverse empirical concerns –  banking,  finance,  gambling, bodies, health, environment, energy, outer-space – are now exploring how to do an anthropology of data that is both committed to describing the richness and complexity of social life whilst also recognising the important role that digital data traces are playing in the formation of this experience.

My main research has focused on how to do an ethnography of climate change and energy in the UK. This is a field replete with data – from scientific data on climate models to home energy data to the data that makes possible the national grid. As I have delved deeper into this data, I have come to ask myself, how might it be possible to bring this data into ethnographic analysis? What light could this shed on energy and climate change as simultaneously cultural and material proesses? What methodological developments would such at attention to this kind of data demand? What kinds of questions should we be asking and what kinds of people should we be asking those questions to? And ultimately what difference might data-ethnography make to our understanding of these complex social/material processes?