Tag Archives: lifestreams

Hiding in Plain Sight

This post draws together some of the many strands that have led to a new project – “Materialising Data, Embodying Climate Change” – which I have just begun with Tom Corby at Central Saint Martin’s, University of the Arts London. I want to make links across what may often seem like a wildly divergent practice – from collaborating with people in the jungle of Papua New Guinea, to working on tech ethics and policy interventions, to building experimental communications platforms, to making abstract sculptures from data, to facilitating workshops in languages I don’t speak with people who have experienced trauma I can barely comprehend.

For me there is a clear trajectory and purpose, which centres on inspiring agency in others. I believe that the answers we need to address our problems – as individuals, as communities, as peoples, as a species – are all around us, hiding in plain sight. The greatest asset we have, as human beings, is our imagination – with it we can devise solutions to whatever we encounter that holds us back, or create new problems for ourselves. Everyone is born with it, and its abundance is the true wealth of our societies and cultures. But while we have a terrible penchant for externalising imagination and corralling it in things – principally valued as commodities – yet it is ever at hand when we need it in challenging circumstances or crises.

Right now we face the implications of our species’ failure to act for a generation in the face of mounting evidence that unconstrained climate change is inevitable and will bring unfathomable disruption to all life on earth. It falls to our imaginations to grapple with these issues – to imagine different paths, to free ourselves from the restraints that have privileged some people’s imaginations and abilities as being more worthwhile or valuable than those of others. I hope with this new project not just to inform, but to inspire people to act.

Materialising Data, Embodying Climate Change is a major three year research project funded by the AHRC which will build on the legacy of the Lifestreams project which I led back in 2012 (with Stefan Kueppers), my 2014 Creativeworks Entrepreneur-in-Residence award with George Roussos at Birkbeck University of London and the experiments with data manifestation and Parkinson’s Disease which Stefan, George and I did in 2016. It also draws heavily on Tom’s long collaboration with Gavin Baily and the British Antarctic Survey creating artworks with climate data, such as Southern Ocean Studies and Northern Polar Studies. It has been a long haul to get here – I first began exploring ideas with Tom in early 2016 around our shared desire to explore how we could engage people with the complex data being generated by multiple interacting climate phenomena (e.g. sea temperature; sea salinity; polar ice extents; methane and other gas levels; krill population etc). For more than a generation, mainstream media coverage of climate change issues have rarely strayed from focusing on just  a single issue – such as the ozone layer, and then CO2 levels – yet it is now clear that it is the interaction of multiple climate phenomena that are driving the changes we seek to control.

I have also had years of discussions with my friend Juan Francisco Salazar at Western Sydney University about how the Lifestreams process could engage with data about the Antarctic. Juan is a filmmaker and anthropologist who has made several trips to the Chilean Antarctic base, and made a feature-length documentary, Nightfall on Gaia, there in 2015. Our discussions have often revolved around how to make the changes that are directly perceptible in more remote and fragile ecological sites, such as Antarctica, accessible in a tangible and tactile way to people living in the urban industrial world, where experiences of the natural world have been muted by human intervention and our connection to nature is fundamentally disrupted. Juan also leads on the Antarctic Cities project, an inspiring collaboration of Australian, New Zealand and Chilean institutions linking three cities (Hobart, Christchurch and Punta Arenas) with each other and Antartica.

The data manifestation trajectory goes back way further too – to a concept of “tangible souvenirs from digital experiences” which I first formulated during the Urban Tapestries project in 2003. At the time I realised that engaging with people from a diverse set of social, educational, cultural and economic circumstances would need the creation of hybrid forms of communication and expression (both digital and physical) which would allow people to engage and participate in ways they are comfortable with. This concept further evolved over the years into experiments with sound and tactile interfaces (e.g. the “Rumbler” and “Sensographs” of  Sensory Threads) as well as paper-based outputs (e.g. automatically generating StoryCubes or DIFFUSION eBooks from digital assets with the bookleteer API).

My experiences working with James Leach and the villagers of Reite in Papua New Guinea on our traditional knowledge documentation project, TKRN (2012-20), have also been of crucial importance in extending my thinking and appreciation not just of the extraordinary range of world views that exist (outside the bubble of Western, industrialised culture), but also of the kinds of knowledge that exist. Particularly those that rely on human senses beyond the visual and aural for meaning-making : such as those of touch, smell, taste, proprioception to name a few. An important insight came to me after my first visit to Reite village in 2012, not long after we had generated the initial set of “lifecharms” or data objects for Lifestreams. I wrote about this in two posts in 2013: Tactile Poetry and Digital Alchemy. James’ writing has also been an influence on my thinking, especially his essays Drum and Voice (2002), Leaving the Magic Out (2012) & The Death of a Drum (2015). My experiences in the village have also exposed me to how direct and perceptible climate change is to people who live within nature and the natural world – seasonal weather patterns do not come reliably; plants are not ready to harvest at the usual times; animals, birds and sea creatures are slowly disappearing. All this is increasing year on year, and is the daily reality for people living traditionally in the forest and on the coast.

Back in September 2017 I wrote up many of my ideas and aspirations for melding these different strands of my work in a post  – Sensing Climate Change Through Empathic Encounters. Almost a year and a half has passed and it seems a good moment to revisit them in the light of my other activities – including developing the UnBias Fairness Toolkit – at the start of this new juncture. All these projects share my key aim of stimulating agency, trust and reciprocity, themes which I have also written about before (Reciprocities of Trust & Reciprocally Entangled) in different contexts, such as my collaborations with Lizzie Coles-Kemp’s Collective Securities group at Royal Holloway University of London and Marina Jirotka’s Human Centred Computing group at the University of Oxford. And lastly, but certainly not least, my long collaboration with Canadian artist and curator Andrew Hunter (since 2007 through his Render and DodoLab projects) and the creative discussions that have sustained both of us through periods of intensive work and fallow, hard times.

The following sections bring together various threads of my work. This essay is a mélange of ideas in development, weaving together the very new and those which have been delicately spun over a number of years. Something which links across all of them is ‘hiding in plain sight’ – the suggestion that what we seek is often already in front of our eyes, but occluded or hidden from our perception. Sometimes because ‘we cannot see the wood for the trees’, sometimes because an obsession with novelty and innovation blinds us to what already works well. For years I have kept a saying of Charles Eames close to my heart – “Innovate as a last resort” – and a copy of the 1953 Eames India Report close to hand. I recommend watching the wonderful film, Goods, in which Charles Eames gives a talk on the manifest joy to be found in the practical simplicity of everyday things – often overlooked but always within reach of our perception.

Engagement and Occlusion: data-veiling

Three years ago, in March 2016, I was invited to present a provocation at a workshop on ethics and responsible innovation in data science hosted by the Alan Turing Institute. It was an opportunity to demonstrate the ideas behind data manifestation to a large (50+) audience of researchers from a variety of disciplines all concerned with ethics in digital innovation. My Lifestreams colleague, Stefan Kueppers, was by then beginning his PhD with George Roussos at Birkbeck and working as a research assistant on the CloudUDPRS project, devising a mobile app to enable Parkinsons patients to self-monitor their motor tests. As part of the trial, we 3D printed a series of lifecharm shells generated from 4 different patients’ data to demonstrate qualitative differences in the individual experiences of Parkinsons that were masked by scoring similarly on the Unified Parkinsons Disease Rating Scale – which is used to determine care & therapy packages. These were given out to the participants as tangible souvenirs.

My talk provoked a number of interesting conversations about the applicability of data manifestation as a means of communicating information on different levels. There were a number of privacy and cybersecurity experts present with whom I began to speculate on the potential for data manifestation (i.e. expressing data in physical forms) as a novel mode of cybersecurity – especially in the field of medical ethics and patient data. I proposed that encoding patient data in physical forms might be a useful means of allowing data to be transported and ‘shared’ with different people – such as a patient discussing their conditions with a physician. In such a situation both parties can interpret the data object and use it to have an informed conversation because they already understand the context in which the patient’s data was collected and from which the data object has been created. However, and crucially, a third party intercepting the data object would simply be unable to access the source data from which the object was generated (e.g. via reverse engineering) or be able to ascertain its context from the object itself. I refer to this approach as “data veiling” – or hiding in plain sight.

My own experience of helping family members and friends through chronic illness necessitating treatment at different hospitals and clinics, illustrated how often – and insecurely – patient data needs to be shared between multiple parties. Often this means patients having to carry paper records of their data with them to every meeting; sending paper copies by post or emailing unencrypted attachments across the internet. The opportunities for private and extremely personal data to leak out into the wider world are rife, constant and a clear threat to individuals given the intense interest in acquiring private medical data now being shown by the companies at the heart of digital technology and AI/machine learning.

Data-Veiling in this way could be an alternative means to achieve cybersecurity – allowing people to carry a physical expression of their data in a way that prevents the source data being directly accessible – the form itself would be enough for meaning to be discussed. It could allow us to discuss with others what the data implies or means by virtue of establishing tactile grammars that help us interpret the shapes and forms. It would remain secure because is unlikely that the source data could be reverse engineered from the physical form, because the (Lifestreams) method of generating the data object is not parametric but a series of structured mutations. The method flows different data streams together in the generation process which cause variations in the growth shape as they interact with each other. This is also notwithstanding additional variations introduced in the 3D printing process itself, such as surface smoothing and slight reductions in resolution depending on the type of material used to fabricate the object – such as metals, plastic, ceramic etc.

Data-veiling also has echoes with work I’ve been doing in Papua New Guinea (PNG) with indigenous people to document traditional knowledge. In such communities what Westerners think of as knowledge is understood and practiced in very different ways; knowledge is often acquired through complex rituals that make manifest personal status within the community and situate a person within a network of relationships. Having and using knowledge is a demonstration of power and ability. In the West, information generally becomes knowledge through its alienation from context into books and other forms of transactable documentation (such as films, digital files etc) that facilitate universal replicability. This is very different to cultures for whom knowledge seems to reside in how relationships are performed between people, place and things. For instance, there may be specific practices (such as magic or sorcery) which cannot be freely shared or discussed openly. Thus documentation (or expression in some kind of object through design) might serve – sufficiently – as a signpost for those who do have the knowledge; it could indicate the lineage that their knowledge has been acquired through and, how transmission to others might occur (through some form of ritual exchange).

In the village I visit in PNG (when referring to practices and phenomena which Westerners would call magic) people slip from speaking in Tok Pisin (the common national language) into their own local language (Nekgini – spoken by less than a thousand locals) and into ‘hap-tok’ (“half speech”) – a kind of allusive mode of speaking around a subject without discussing it directly. The ability to participate in and understand the meaning behind such discussions would demonstrate either knowledge itself, and the ability to understand how to ‘read’ the signposts being referred to, or to the constraints under which such knowledge could be acquired from those who have it. Porer Nombo, one of the village elders who James has worked with for over two decades, suggested something along these lines when giving a presentation about the ethno-botanical book he co-wrote with James, Reite Plants (as reported in James’ essay, Leaving the Magic Out). To me, this is another kind of hiding in plain sight – documenting part of a story and including clues for the willing to discover how further layers might be accessed. Revealing just enough, sufficient for others to make use of what is spoken.

By encoding sensitive data into physical abstractions it might be possible to conduct informed discussions about the real world meaning of datasets within specific contexts – without having to access or share the data itself all the time. Data veiling could offer a form of signposting or symbolically representing sufficient data without having to reveal its full detail. Shifting our focus from data to the patterns and meanings we can interpret from it.

Reciprocity, Care and Safety

Over the past two and a half years my attention has been largely taken up with my role in the UnBias project, and in particular, devising practical and pragmatic ways to foster both an awareness of bias, trust and fairness in algorithmic systems, and how to “do ethics”. ‘AI Ethics’ went mainstream in 2018 and there is now huge interest in how those companies which build digital technologies and services (especially those involved in AI and machine learning) can deliver responsible innovation through ethical design and development processes.

The egregious harms to democracy, as well as to individuals, which have been exposed in the past few years (Snowden, Cambridge Analytica etc) give this a tremendous urgency. More and more systems and decisions seem to be being automated, with all the damaging effects long-predicted becoming day-to-day reality. Whilst human consciousness is able to deal with contradictory states and can make informed judgements that reflect the complexity of contexts and situations, we are yet to devise machines that can handle such complexity themselves. Perhaps this is an inherent weakness of the binary mathematics they are built on? Increasingly the evidence emerges not only of systemic bias being found in automated systems (mostly trained on inherently biased datasets), but also of how such systems are amplifying pre-existing biases and prejudicial outcomes which have the greatest impact on the most vulnerable in society. I fear that the kinds of automation offered by AI/Machine learning, based essentially on inferences drawn from incomplete and prejudicial datasets, might only ever be a zero-sum game that inevitably leads to a “statistical… regression to the mean” (quoted from Alan Blackwell, below).

The danger is not the creation of systems that become maliciously intelligent, but of systems that are designed to be inhumane through neglect of the individual, social and political consequences of technical decisions.
Alan Blackwell, Interacting with an Inferred World: The Challenge of Machine Learning for Humane Computer Interaction

The mainstream narrative for increasing automation has for a long time revolved around “efficiency” – that machines are more efficient than people, they cost less, are somehow ‘neutral’ etc etc. We need to closely interrogate such narratives and expose their fallacy for what it is: a mask to cover the reality of where that vast bulk of the benefits flowing from such automation go. This is hardly a new situation – for hundreds of years people whose ways of life and agency have been undermined and eroded by capital and automation have tried to resist what is essentially a subjugation. Their descendants created unions, cooperatives and formed mutual societies to win back some of their lost agency and initiative, but history shows us it is a recurring theme, for which almost each generation has to find new answers and approaches.

We are also beginning to see the end of another narrative that has specifically woven itself around digital technologies – that innovation should be unfettered by regulation, and that the digital tech industries are creating change faster than laws and regulations can keep up with them. We live in a time where a corporatist agenda seems to have gained a triumphal ascendancy, and where mantras like “information wants to be free” are posed against a picture of a heavy-handed state endlessly creating red-tape to frustrate the innovators and entrepreneurs who are cast as the only ones who create new value in society. Such narratives are as hollow and self-serving as those which reduce all values to the purely monetary and refuse to account for key human values such as kindness, love, generosity, happiness, play, pleasure and joy.

Returning the the theme of hiding in plain sight, I have drawn great inspiration from the work of William Perrin and Lorna Woods on the concept of duty of care in social media regulation for the Carnegie UK Trust. They look back to the groundbreaking UK Health And Safety legislation of 1974 which harmonised protections for workers across all industries and workplaces under a general duty of care, and extended those duties to pro-active harm anticipation and reduction through due diligence to the rest of society. Their suggestion is to build a new regulatory framework which draws upon the well-established principles of Health and Safety legislation to create a proactive duty of care on providers of digital services and platforms for harm reduction. This would shift the emphasis from the current post hoc situation where liability is assessed and punished only after harm has been experienced, to an a priori approach based on taking care as a first principle. A fundamentally diametric approach to the infamous Silicon Valley mantra of “move fast and break things”. Digital Safety rather than Digital Security – proactive efforts that care for rather than seek redress for harms already done.

What I like so much about this approach is that it builds trust based on reciprocity – when we know that people are exercising a duty of care on all our behalfs, we trust that they will do their best to not harm others. Trust is both formed through, and builds on, relationships and the performance of the values that underpin them, not simply through transactions. For this reason I remain highly skeptical of the utopian claims for ‘smart contracts‘ that have surrounded the blockchain hype in recent years. Whilst they may offer some benefits in terms of automating certain types of exchange, I see specific dangers in attempting to reduce all kinds of complex relationships and exchanges to contractual standards based on transactions. What of trust? What of reciprocal obligations and fundamental rights? There is always a tension between human rights and contract law – an asymmetry where the weaker party is often induced to sign away their rights for incommensurate benefits to the advantage of the stronger party. Unless we have robust institutions and frameworks for the protection of the vulnerable, and people have access to education and information about their rights and values, we will forever be at the mercy of those who corral power and wealth for themselves, and seek to buy their way out of obligations to others through unbalanced contracts and negotiating power.

Who gains from a world where we are increasingly encompassed by contracts that privilege transactional relationships? In which social domain are such contracts disputed and settled? Traditionally lawyers and those who can afford to be litigious have been the gatekeepers and prime movers in contractual situations. Smart contracts may seek to subvert this nexus, but will they really democratise these processes or simply alter the landscape of who holds the upper hand in favour of the programmers – and their employers – who are able to understand and manipulate the arcane new languages that evolve with them?

It seems to me that human rights and concepts like duty of care and the precautionary principle are a critical baseline that are collective in nature – i.e they support everyone in society regardless of status or situation. In contrast to this are contracts and transactional systems which inevitably privilege negotiations for those already invested with power and resources. If we truly want a fair and just society, supported by tools and technologies that we trust, then it seems to me that we must reinvigorate constitutional democracy with transparent rights and laws, as well as accessible education and information about what they mean and how they affect people. That way we can take care of each other.

Inspiring and Stimulating Others

Despite the focus of my new project on climate change, I’ve barely touched on it here and it is, to my mind, the most critical issue we face as a species living in a fragile ecology. My preoccupation with agency is, though, at the very heart of how I think we can inspire and stimulate the kinds of societal change necessary to preserve as much of life on earth as we can. I am neither an ecologist nor a climate activist, but I believe that each person must find within their own skillset and area of work and life the means to contribute positively to such change. Not just in terms of our personal habits of consumption, but in the effect that our work can have on others and the processes we live within. As an artist and designer, I feel it is my duty to achieve this the best way I can – by inspiring others and, ideally, stimulating them to be hopeful and to take agency for themselves.

I have also been inspired by coming across a remarkable initiative that took place in the UK in the 1970s – the Lucas Plan. Faced with the threat of massive redundancies the shop stewards of several unions representing workers at Lucas Aerospace formed a ‘Combine’ to develop an alternative business and industrial plan for the company. This was based on producing ‘socially useful products’ (at the time 50% of Lucas’ products were parts for military hardware) and around 150 such products were proposed by the workers, based on existing skills, tooling and capabilities. Ultimately they were dismissed by the management board. Never-the-less, the initiative inspired many other outcomes and initiatives – some of which continue to this day. (Incidentally, many of the then experimental products are now mainstream). Watch a trailer for Steve Sprung’s forthcoming documentary, The Plan, or an original 1978 documentary by the Open University/BBC. To me this is an example of what happens when people begin to take agency for themselves and reimagine some of the basic premises about what it means to be responsible for their actions – for the things they produce as a worker in a firm or institution, for the impact this has on society and the environment. With this sort of thinking and action we might yet reconfigure our society and industry towards a trajectory that minimises the impact on the ecosystem and life in general, and mitigates some of the worst excesses of the past for a sustainable future.

One of the key figures in the Combine was Mike Cooley, whose 1980 book, Architect or Bee? The Human Price of Technology, sets out a clear vision for human-centred, socially useful design and production. His thoughts on ‘augmented and symbiotic systems’ in favour of automation or classical AI systems are particularly brilliant. He proposes that such technologies be created to enhance human capabilities, not replace them. He describes the path to expertise as being one which increasingly expands on the basic rules acquired along the learning journey adding intuition, insight and imagination as one’s experience grows. The goal he suggests is for technologies to enable and augment everyone’s capacities, not to de-skill people merely to increase the profit share for managers and owners. He calls for humane technologies that not only increase our abilities to make things, but to appreciate life, freedom and choice as well. In one example, Cooley writes of designing expert medical systems in the 1980s which,
provide an interaction between the ‘facts of the domain’ and the fuzzy reasoning, tacit knowledge, imagination and heuristics of the expert, and no attempt is made to reduce all these aspects to a rule-based system – the system is seen as something that aids rather than replaces the expert.

Although not long, the scope of this book is vast and is full of insights into the operation of power in society, politics, business and industry as well as offering practical suggestions for how to address it creatively and with intelligence. It is a vision for empowering people to act for mutual benefit, in sympathy with the ecosystem which sustains us. As we grapple with the choices ahead of us for democracy and ecological sustainability, we could do worse than take a few leaves from this book:
The choices are essentially political and ideological rather than technological. As we design technological systems, we are in fact designing a set of social relationships, and as we question those social relationships and attempt to design systems differently, we are then beginning to challenge, in a fundamental way, power structures in society.

My hope for the Materialising Data project is that we can find new ways to create empathic encounters with the urgent, complex reality of climate change for people to whom it is not yet as directly appreciable as it is for those living closer to nature. In challenging the dominant modes and forms of how we make meaning from data, I hope that we too can begin to challenge some of the power structures in society that have remained inert in the face of growing emergency for almost my whole lifetime. We have a duty of care for future generations, and for all the other forms of life we share this planet with. The answers remain before us, hiding in plain sight – we just have to be brave enough to adopt them.

Despair is not an option:
Anticipate the worst;
Hope for the best;
Cope with what comes;
Tread lightly on the earth.

London, February 2019

Data Manifestation Talk at ODI

Video of my talk at the Open Data Institute on Friday 17th June 2016:

For further details, read my post, How Do We Know?

How Do We Know?

The following is the text of a ‘provocation’ I gave at the Responsible, Ethics-Aware Research and Innovation in Data Science symposium hosted by the Alan Turing Institute in March 2016. For the event, my colleague Stefan Kueppers and Professor George Roussos at Birkbeck University of London generated a new set of Lifecharm shells using data from a research trial with sufferers of Parkinsons Disease. We 3D printed multiple copies of 4 shells which were given to the participants as a tangible souvenir of the data manifestation concept.

**********

I’m going to speak today about a concept I call “data manifestation” and how I believe it can add a significant dimension to data science and how we make meaning from data.

First I wish to pose a question –

How do we know what we know?
And how do we make meaning from what we know?

The answer is, of course, through the interaction of our senses with the stimuli coming in to them from the external world – and how this is processed by our consciousness and woven in together with our memories and emotions.
This is a system of high complexity and entanglement. It produces judgements based on multiple factors and dimensions all working together.

I believe that this is qualitatively different to the kinds of inference driven systems that are increasingly coming to dominate our civilisation and how it operates.

For instance: when we feel that something we have eaten is delicious, how do we know this?
It is not just the taste that defines this judgement, but includes many other factors such as smell, look, texture, sound, shape and size, whether it invokes memories, or causes a momentary shift in our sense of balance. If each of these factors is seen as a data input, how would we begin to map such complexity into the decision-making systems of today?

Art and aesthetics are another example of this complexity. When you encounter a work of art the experience itself which determines your aesthetic reaction to the piece – whether one of awe, delight, revulsion or indifference – is driven by similar complex factors. The use of materials, colour, scale, lighting, etc as well as memory and emotion create each person’s own aesthetic experience. There can be no right or wrong aesthetic experience, only how each individual experiences a work of art in relation to the summation of their own existence.

Much of the data that is captured, recorded, analysed and used as the basis of inferential decision making systems in the everyday world is principally derived from visual interfaces, from the writing of the software that generates it to the representation of the data in screen-based visualisations – from spreadsheets, to graphs and animations. There are in a few cases some sonic representations of data, and in extremely rare cases some haptic interfaces. But on the whole the way we create, analyse and present data is visual and screen-based.

But not all experiences are visual. And my proposition is that by failing to encompass the whole and extraordinary sensitivity and careful tuning of the human sensorium in our methods of expressing data we are missing huge potential.

Back in 2012 Proboscis was commissioned to collaborate with scientists at Philips Research Lab in Cambridge as part of a public art programme. We were asked to explore the problem posed by the failure of commercial biosensor devices to engage people in leading more healthy lifestyles. Essentially the FitBits, Fuelbands and similar devices were being switched off within a few weeks of activation. Philips is a big player in the separate business of TeleHealth (where people with serious medical conditions have instrumented homes that record and transmit data to health professionals and back to the users). They were interested in how these emerging life-tracker devices could have positive well-being and health benefits for nominally healthy people over the long term. But the usage data suggested this was simply not happening.

Looking at the problem, our insight as artists was that this was an issue of relevance to people’s sense of their own self, to their identity. Our hunch was that the ways in which the narrative of a person’s life helps construct their sense of self were just not being engaged by the graphs of data represented on smartphone screens or laptops. Humans frequently invest meaning in objects that act as triggers for memory and personal significance. These can be highly ritualised objects such as wedding rings or other kinds of jewellery. They can also be seemingly random objects, such as a piece of cloth or a pebble from a beach. With all these things, the memory is evoked not just by the sight of the object, but through other factors such as the way it feels.

Our insight was to connect this tactile relationship to meaning with the kinds of data that could be collected by tracking devices. We built a simple data logger with off the shelf components and collected a range of data over a week from members of our studio: step count; pulse rate, sleep patterns, blood pressure, stress factors etc.

My colleague Stefan Kueppers then developed an algorithm based process to flow this data into a shape that could be 3D printed. We looked to nature for inspiration and realised that the mathematics and geometries in shells would allow for extraordinarily complex and unique expressions of data in a 3 dimensional form. Here are a few examples of the objects we made from our own biodata. We called these “Lifecharms” and the project, “Lifestreams”.

What I am going to show you next is our latest experiment. We have been collaborating with Professor George Roussos at Birkbeck (where Stefan is currently working on his PhD about the Lifestreams work) to adapt and extend the data manifestation concept to other uses.

Birkbeck have been running a data collection project working with Parkinson’s Disease patients – mapping the wide variation of symptoms that they experience. Parkinsons is a disease encompassing high variability in symptoms between individuals; ideally this ought to imply a high degree of personalisation in their medical care and disease management regimes. Yet health policy and medical care management for Parkinson’s sufferers are determined by a ‘scale’ (called the UPDRS) derived from about 70 factors of motor performance criteria a Parkinsons patient is classified against. These factors are collapsed into a single “summary statistic”, which is then used to assign treatment for patients according to where they lie on the scale. Whilst this is an effective way to communicate the multiple dimensions of Parkinson’s on a linear scale of progression, because of the enormous variation it is common for people with entirely different symptoms to score similarly on the scale, yet require very different care.

For patients and health professionals alike, finding ways to express the unique characteristics of an individual’s actual experience of Parkinson’s would be a big step both for personal dignity and as a way to convey the wide variations in an easily appreciable manner to policymakers.

What you see in these four examples here, a version of which you are being given to take away and keep, is an expression of three data sources from four individuals who participated in the Parkinson’s trial, each mapped in 3 dimensions into the form of a shell. They demonstrate the clear variations in symptoms experienced by these individuals. The data used were :

1) Tremor in the Left Hand – which maps to how much the spiral of the shell stretches outwards.

2) Left Leg Agility – which maps to the overall growth scale of the shell (i.e. how big it gets in terms of volume)

3) Two Target Fingertapping (Left Hand) – which maps to the change in the frequency of the recurring ridges on the exterior of the shell.

These four shells are only a very initial experiment and there are many more factors that we will be able to control with data sources from the trial in future iterations, such as the number and increment of rotations, ripples in the curve of the shell exterior, the sweep curve of the main shell, twists in the shell curve, other geometric entities occuring on the shell surface (spines and nodules) as well as the number of segments that make up the whole shell.

A key indicator in Parkinsons is how it affects a person’s left and right sides of their body so, in the future, we also propose to develop twin or “clamshell” forms that spiral out from a common centre and reflect differences recorded between a person’s left and right symmetries.

As you can see, and soon feel, even with just 3 data sources informing the shell growth parameters, it is possible to appreciate the individuality of each person in a way that is qualitatively different to looking at a graph. As we introduce more granularity and complexity, the individuation will become even more pronounced.

Taken as snapshots, such shells, could over time reflect the changes in Parkinson’s symptoms as experienced by sufferers.

What use is this? We believe that this kind of approach offers a very different and rich way to express complex datasets for making meaning in a world where data is ubiquitous, is increasing exponentially and – as we are constantly being told – is overwhelming us.

My proposition is that in this way we can find alternative ways to make meaning and take decisions based on human insight and judgement from complex data sets. We don’t just have to simplify and summarise data in linear ways to make it easier to represent visually on a screen, we can also use our other senses – touch, sense of scale and balance, taste, smell, hearing, time and temperature. We can also benefit from other knowledge traditions (as well as contemporary science) and apply methodologies and critical analysis from the arts and humanities, such as aesthetics, to make meaning and draw judgements from highly detailed material artefacts that reify complex data sets into actual things, not just representations on a screen.

Thank you.

Characteristics of the Shells

Shell 1 – spiral not so pronounced, which indicates that the left hand tremor amplitude is not so strong; the shell size is quite big indicating the patient has significant increase in amplitude of left leg tremor; the ridges are also become quite pronounced indicating that there is a growng lack of accuracy and slowness in finger tapping.

Shell 2 – the spiral is at the higher end, suggesting the patient has a higher amplitude of left hand tremor; the shell size is a bit smaller so the patient experiences less left leg tremor amplitude; the ridges are very small indicating a higher degree of accuracy and speed in the patient’s finger tapping.

Shell 3 – the spiral is a median, indicating a moderate but not pronounced amplitude of left hand tremor; the shell size is slighty above average, suggesting increasing tremor in the left leg; and the ridges are fairly pronounced suggesting a greater lack of accuracy and slowness in finger tapping.

Shell 4 – very pronounced spiral suggesting very high amplitude of left hand tremor; shell size is smaller indicating less amplitude of left leg tremor; the ridges are very pronounced suggesting a high lack of accuracy and slowness in finger tapping.

Embodying Data Workshop

Last week I was in Edinburgh to run my co-discovery Embodying Data Workshop with 24 of Chris Speed‘s Design Informatics Masters students at the Art College. The workshop was devised a few months ago as part of my Creativeworks residency at Birkbeck College to introduce computer scientists to the possibilities of approaching the problem of data analysis and computation differently by manifesting data in tangible ways. Thus we may bring more of our human senses to bear on meaning making than merely relying, almost exclusively, on vision and hearing as with standard data visualisation techniques.

It is a hands-on workshop using paper-prototyping methods to explore manifesting data as physical objects – why would we do this? what sort of objects? how could it benefit data analysis and computation methods? – and is an opportunity to speculate on how we might discover new ways to generate insights into complex data sets to discern previously undetected patterns and make meaning.

The workshop starts off by engaging the participants in identifying 11 major human senses (touch, taste, smell, sight, hearing, temperature, balance, pain, time, proprioception and interception) and discussing other sensory factors. Then we begin to map how many senses we actually use in everyday activities and tasks. In this way it becomes apparent how reliant we are on multiples senses to interpret our experiences of the world around us. How much then are we missing in trying to analyse data using just our visual and occasionally auditory senses? What patterns might also exist that we are simply not able to perceive because the senses that would detect them are not being activated? The last part of this exercise asks the participants to think about data sources and types, such as different kinds of sensors or data feeds/streams as well as the kinds of data coming through – energy use, health, environmental, sales, traffic, communications etc. We then follow on by mapping how we might interrogate such data using additional senses to sight and sound. What benefits might we get from having new ways to explore big and complex data sets? What could happen when we take digital data out of the machine and into the physical world?

IMG_2891 IMG_2892 IMG_2893 IMG_2894

IMG_2902 IMG_2903 IMG_2904 IMG_2905

This exercise is followed by a short presentation on Proboscis’ Lifestreams project: how and why we developed the life charms from biosensor data and what implications we believe it has for creating new insights into health and fitness data for wellbeing. Lifestreams provides both an actual context and talking point to discuss the difference that embodying data in the physical world – making it tangible to the senses – could have.

The second workshop exercise involves each participant using a blank StoryCube to imagine a data object of their own. I encourage them to use the 6 sides of the cube to indicate data types or streams that they might be using in their existing work or projects which could be used to generate a hypothetical data object. Then, placing their cube on a worksheet, the participants are prompted to consider how their data objects would connect with different human senses; how the objects might interact or connect with each other; whether they are personal or shared objects; what kind of conditions might people encounter them in. Participants are encouraged to consider what implications may arise from all these too.

IMG_2895 IMG_2896 IMG_2897 IMG_2898 IMG_2899 IMG_2901

Finally we discuss the ideas that have emerged from the exercise and, more generally, the potential for new insights into complex datasets to emerge from creating the possibility for senses other than sight and hearing to be involved in analysis. The point of the co-discovery workshop is to allow participants to come to their own understanding of the potential and what might be possible, not to be didactic. It aims to plant a seed of curiosity by exploring the gaps and absences in our toolkits for creating new kinds of knowledge, hopefully to inspire entirely new ways of expressing data in physical or environmental forms such that we can move beyond the ghetto of the primacy of the screen.

Further Iterations
I am available to deliver the workshop to other groups (academic, cultural or corporate). Please contact me for prices and bookings. The workshop lasts about 3 hours, is suitable for a range of abilities and works best with groups of 10 and more. It has been designed to engage scientists at both postgraduate student (Masters/PhD) and professional staff level (lecturer/researcher). It works equally well with designers, artists and others who are also exploring the use of data as a creative material/medium.