Data versus stories, abstract models versus everyday complex
Ed. This is the quite different, original version of the foreword written for After the Flood’s ‘Cities Squared: Making Urban Data Legible’ publication, about their data viz tool for London, London Squared. I was fortunate enough to be able to commisson the original work from After the Flood, as explained here, as well as to collaborate with them on it a bit, though it’s really all their work—as is this book. ‘Cities Squared’ extends the work well beyond London’s footprint, into other cities and other ideas. The book is predictably solid, useful, and desirable, as well as insightful, and it’s particularly rewarding to be able to ‘close the loop’ by writing for their publication about the broader context of data about cities, models of cities, and understanding and visualising cities.
The statistician George Box famously said “All models are wrong, but some are useful”. It’s a less-than-useful statement, no matter how accurate, recalling nineteenth century Philadelphia retailer John Wanamaker’s, when he said that half of the money he spent on advertising was useless — he just didn’t know which half. But Box’s adage does capture the maddening pull of modelling and mapping and visualising: a beguiling sense that we can capture the complexity of the world around us and reduce it down to some essential meaning, to be understood, calibrated, controlled.
And in an age increasingly defined by three interweaving patterns which apparently lend themselves to modelling and mapping — the dominance of the city, the technologies of the Internet, and the climate crisis — it is hardly surprising that modelling, mapping and visualisation seem so vital. The ongoing fall and fall of the smart cities movement notwithstanding, the idea that we can visualise comprehensive urban data dominates technical conversations in city halls across the world.
It’s a fundamental analytical shift from the station to the taxi, from the fixed dimensions of the street to the fluid interactions of the sidewalk. The form of data changes, and the tools and capabilities accordingly.
Currently, these conversations are obsessing over the idea of the ‘digital twin’, an attempt to model and manage a city through its data, real-time and otherwise, about almost everything, mobile and static. The Twin seems possible due to the now-usual bundle of Internet-of-Things tech, machine intelligence, cloud connectivity and data viz, but thrown at our cities’ streets, and their own heap of disparate components; air quality, scooters, retail spend, spoken language, autonomous vehicles, parking spaces, whatever.
Within cities, this moves management from the largely static, inert structures of the last couple of centuries — a paper ledger could contain much of the data required to map fixed assets such as libraries, factories, and freeways, in that sense — to these more fluid arena of mobility, energy, consumption, even migration.
It’s a fundamental analytical shift from the station to the taxi, from the fixed dimensions of the street to the fluid interactions of the sidewalk. The form of data changes, and the tools and capabilities accordingly.
Mobility in particular has driven this shift of focus. From Stockholm to Shanghai, growing cities are producing fast-moving mobility demands well beyond that which their slow-moving physical form can support, whilst simultaneously being seen as a proving ground for a heavily-funded tech sector in ‘blitzscaling’ mode. Hence a pressing need to pull focus on transient, ephemeral flows of people and things and spaces and qualities, rather than the inert stone structures they sit within or move upon. The digital twin here offers what CityLab calls “a virtual replica of real-world mobility flows” (whilst also noting the accompanying privacy and capability issues.) Whereas a typically hyperbolic Forbes prefers to see the Twin as all up-side:
“Digital twins are powerful masterminds to drive innovation and performance. Imagine it as your most talented product technicians with the most advanced monitoring, analytical, and predictive capabilities at their fingertips.”
Great! Yet there remain some nagging issues: do Twins replace ‘talented product technicians’ or augment them? What or who are these ‘talented product technicians’ in cities? Can Twins truly capture the complexity of cities? And so just how much faith should we place in data?
Data is given as a starting point
Data, in itself, is not something to place faith in at all; it is something to work with. In ‘Being Ecological’, the philosopher Timothy Morton writes that the word ‘data’ essentially means ‘given things’. ‘Data’ is the plural form of the Latin datum, ‘that which is given’, from the Latin verb ‘to give’. In other words, data is not the self-evident truth of an object, or a fact, even. It is something we are given to work with, something we have to interpret, more akin to a complex process than an entry in a field.
“In order to have a fact, you need two things: data, and an interpretation of that data … Common talk imagines facts to be things like barcodes that you can read off a thing: they are self-evident. But a scientific fact isn’t self-evident. That’s precisely why you have to do an experiment, collect data and interpret that data.” (Timothy Morton)
This emphasis on experiments, of obtaining data and working with it to interpret it, is useful. It shows us that data is essentially more subjective than the word is usually taken to mean. This is not a weakness. Rather, it is its strength, in that it emphasises the work to be done, in terms of engaging with the world, in terms of understanding context, seeking out what is hidden as well as obvious, contrasting with other knowledge, and drawing out the real limits of data.
The glimpses of a city’s data that Twins enable, when visualised, are only the start of a process of ongoing understanding, which actually unfolds in public, and through interaction with the street. That glimpse needs to be worked through, developed, and augmented with other practices. Given the public context, this requires multidisciplinary municipal teams to the fore, including designers as well as data scientists, artists as well as analysts, sociologists as well as software engineers. Such a capability could properly dig into the data, as Morton suggests. Otherwise they merely represent a limited theory about the city, and as the philosopher Graham Harman says, “no theory survives its first contact with reality.”
Or, as Mike Tyson put it, “Everybody has a plan until they get punched in the mouth.”
But the glimpse is a useful token for unlocking this practice of everyday life nonetheless. It glimmers with the promise of understanding. The glimpse is the start of something.
John Le Carré said the only way to write about a place was after visiting it for a day, or after a long life once you’d moved there. The time between those two lengths was somewhat useless, in terms of understanding. It didn’t lend more clarity; it just accumulated detail. Le Carré’s first day in the city is another glimpse, and can actually be profoundly insightful about a place.
But equally, the danger of the glimpse in another context — the high-level data visualisation of the Twin, in pressured conditions, or in the wrong hands – is that it will falsely present as something deeper. The next level of insight for Le Carré was only unlocked after a long life lived in a place, yet few have such patience in contemporary politics and business, and they may rush into assuming they have reached their destination.
In this sense, the journey is more important than the destination. It is this spirit we should bring to mapping and understanding the city — as an ongoing process, to be constantly reviewed from multiple viewpoints, and via multiple formats, the glimpses in-between Le Carré’s first day and long-life, but then the working through of data in practice, in context.
The toolkit described in this book is self-aware enough to position its abstraction at the right level. It is clearly not enough in itself; yet it is an accessible visual hook to hang a conversation upon, subtly drawing the viewer into the data. As such, it can sit humbly alongside more deliberately subjective approaches, equally valid ways of representing the city, no matter how personal.
This more actively engaged mode also means we are likely to encounter many different perspectives on the city, beyond the simplistic data of the Twins. Even an apparently static narrative form such as street photography provides rich insights into cities; not least because, as Geoff Dyer described, it is actually a constantly moving practice. Dyer described photography as “the ongoing moment”, modifying Henri Cartier-Bresson’s prior definition as “the decisive moment.” City officials and tech companies would no doubt love to claim that city models and maps show the ‘decisive’ elements of cities, yet they can really only be part of an ‘ongoing’ process. As long as we recognise this, we have no problem.
To get a glimpse of Tokyo, would we only visualise real-time public transport data, as revealing as that would be … or could we also counterpoint that data with the street photography of Daido Moriyama or science fiction of Yoko Towada? Or draw historical insight from the carefully researched manga of Hinako Sugiura or its opposite from an episode of Tokyo Ghoul? Or from the conversations stimulated by chunks of architecture dropped into Jiyugaoka? All are valid readings of that city, across past, present and future, and there are almost infinitely more out there like that.
Or, as the Swedish crime writer Per Wahlöö, who wrote the influential Martin Beck series with partner Maj Sjöwall, described, we might instead:
“Use the crime novel as a scalpel to cut up a society… and simply find who is responsible for what, and if there is anything left to be responsible for.”
Wahlöö and Sjöwall’s books, leading directly to Wallander, Miss Smilla, and ‘The Girl with the Dragon Tattoo’, continue to shape perceptions of Swedish cities decades later, rightly or wrongly, because the cities we inhabit are really what Jonathan Raban called ‘soft cities’, composed of subjective perceptions, evanescent social contexts, ephemeral events, cultural expression, misunderstandings and recombinations — put simply, they are created by what you bring to them — as much as they are assemblages of hard infrastructure, hard geology and hard data.
“(The city) invites you to remake it, to consolidate it into an shape you can live in . . . Decide who you are, and the city will again assume a fixed form around you.” (Jonathan Raban)
No wonder cities are difficult to pin down with a few lines of code.
And there is something else in the Digital Twin rhetoric that should also give us pause: that other kind of twin, the doppelgänger.
Today, the doppelgänger tends to imply an ominous sensation, often a spectral ‘evil twin’. (And so a staple of many a horror movie, from Kubrick’s ‘The Shining’ to Jordan Peele’s recent ‘Us’.)
The German word, and the ominous meaning, is a relatively recent, though; the concept itself has a longer history, and is perhaps more even-handed. The version in Finnish mythology, the ‘etiäinen’, is a spirit that all places, things and human beings have, an image or impression that goes ahead of a person, doing things the person in question later does, essentially a kind of déjà vu in reverse. It lends a peculiar feeling that something is about to happen, and yes, in a typically Finnish mode perhaps, that a bad year is about come. But not necessarily. As an apparition, it is a kind of working experiential model of the future. This ‘etiäinen’ is itself a local variant of the ‘vardøger’ in other Norse mythologies, another kind of a spirit predecessor, (an instance of the broader guardian spirit, ‘fylgja’) which similarly appears as a glimpse of the future around the corner, benevolent or otherwise. (The concept also pops up in Ancient Egypt — the ‘ka’, who pops up in Euripides’s play ‘Helen’, to mislead Paris — and old Irish folklore, with the ‘fetch’, another one of the darker variations on this theme.)
All this suggests a deep-seated longing for the foresight that such a twin, in the form of a model of the self, could bring us. These patterns in mythology, all frustratingly supernatural glimpses leaving humans firmly in their place, have now been translated into today’s anthroprocentric belief systems, and so they are rendered in arrays of frankly less alluring LED displays, simplistic scatter plots, and the promise of ‘scientific’ knowledge from real-time data (and recall again Morton’s warning about that).
Yet when attempting to sell its version of the Twin, IBM says “the advent of digital twins offers engineers a technological leap ‘through the looking glass’ into the very heart of their physical assets. Digital twins give us a glimpse into what is happening, or what can happen, with physical assets now and far into the future.”
It’s interesting that IBM uses this phrasing of the “glimpse”, as well as the allusion to a near-magical sensation, as if etiäinen. So perhaps this long-standing uneasiness about Twins should also be recognised. There’s something to it.
Equally, though, it reinforces the idea of the glimpse as far-from-perfect knowledge, but as providing a hint of what could be around the corner, depending on what we do next, how we grapple with the complexity of taking action.
The digital version of the Twin, and much of the visual language of urban data visualisation, has emerged, subconsciously or otherwise, from a generation that grew up playing the computer game ‘SimCity’ (1989). Whilst the game itself functioned as kind of ‘glimpse’ into the world of urban policy and planning, it’s a safe bet that few of that generation unpicked the political backdrop that invisibly defined the game’s mechanics. Will Wright, the game’s designer, was no urban planner; he drew most of his understanding of how cities work from a 1969 book by Jay Wright Forrester called ‘Urban Dynamics’ (This story is detailed brilliantly by Kevin T. Baker for Logic.)
Forrester is a divisive figure in the previously niche subculture of urban modelling and simulation, but hugely influential more broadly in systems theory and computer science. In 1968, whilst working at MIT, Forrester claimed he had finally reduced the problems of the city to 200 parameters arranged over 150 equations. One could imagine the attraction for the politicians of the day, and his theories took flight under influential Nixon adviser Daniel Patrick Moynihan.
Those theories were not ‘simply technical’ at all; rather, they were politics dressed up as research. In essence, Forrester’s belief that complex social systems are actually beyond human comprehension led him to conclude that any social policy that attempts to improve conditions will inevitably backfire, usually making things worse. Under Nixon’s administration, and arguably many since, that meant concluding, for instance, that social housing programs can only result in condemning the unemployed to lifelong poverty, no matter how well-meaning. This is patently false, as any number of successful social housing schemes elsewhere indicate. The conclusion entirely depends on the political ideologies framing the choice of variables and algorithms, rather than any inherent truth. Yet similar policy principles to Forrester’s are buried deep in the code of SimCity, and manifest in the game’s emphasis on calibrating a few decisive economic growth variables. Despite ignoring the true complexity of cities, SimCity still subconsciously shapes much of the language and colour now surrounding ‘digital twins’, city models and urban data. Note that both Forbes and IBM suggest the users of twins are ‘product technicians’ and ‘engineers’, a telling choice of disciplines.
(See also today’s inheritor of SimCity’s crown, Cities:Skylines, which also tends to preference the misleadingly easier variables in cities, appealing to technocratic viewpoints. In the words of Youtuber donoteat01, aka Justin Roczniak, “Free-flowing traffic is prioritised above all in the virtual world (of Cities:Skylines) even more than it is in real life.” Roczniak’s critiques are well worth watching—and gain bonus points for adapting the cover of Robert Caro’s ‘The Power Broker’ for their poster image.)
Ironically, Forrester’s sense that ‘complexity’ is too, well, complex for social policy meant that it was also too complex for his urban simulations, which had no room for questions of politics or conflict — in other words, for people. (Just as SimCity is a hopelessly simplified representation of ecosystems, social or otherwise). The danger now is that we pretend that our models will now, finally, work. That they just got better. That we are simply just much smarter now. Indeed the models may be better, technically — Forrester was working with punched cards and teletype printers, after all, and now we have (pause for choral interlude) “…AI!” — but they will always fall short of the necessary reading of reality that a social housing programme necessitates, for example. That can only be handled by combining data with interaction, action and representation — ‘showing up’, in other words, in the complexity of the everyday.
Further, that increase in system complexity can also render policy illegible. When Kevin Slavin says “we are writing code that we cannot read”, this is problematic enough when those algorithms apply to ad-tracking and recommendations engines. But when applied to city services like transport, housing, energy, waste, water, and urban planning more broadly – decision-making which has previously been conveyed via accessible and legible public record – it is more fundamentally troubling again.
Of course, the metaphor of ‘reduction’ and collapsing down complexity is appealing in a sense. You can understand the particular appeal to time-pressed, under-resourced city officials with social and other media at their back, and politicians breathing down their neck — but is remains a false goal, if that is all it is. A reduction actually remains complex, and still needs combining with other approaches.
Learning from the idea of ‘a reduction’ from cookery, a discipline that actually practices complexity everyday, it is clear that a sauce is something other than simply the sum of the parts. (Try telling Auguste Escoffier that the five Sauces Mere in his 1902 ‘Guide Cuilinaire’ are simply the sum of their ingredients. Or that the sauce is all we need to make a dinner.) The reduction is another thing entirely, albeit interconnected and interdependent, and remains complex is its own right. We cannot tame complexity; only wallow in it.
This is not a bad thing, no matter how confronting to orthodox management theory. Rebecca Solnit wrote that “the magic of the street is the mingling of the errand and the epiphany”. This also gives us a clue that it’s the ongoing everyday mundane as well as the “decisive moment” that we need to understand. And we are surrounded by such magic in cities.
The writer Gautam Bahn, reflecting on the failure of India’s smart city programme, wonderfully describes “the survival of the ordinary and the everyday; the survival of citizens over cities; of infrastructures of everyday dignity over big, signature, spectacular projects; of incremental change over instantaneous transformation; of the bazaar over the mall, the shared auto over the expressway”
We can easily comprehend what I call ‘everyday complex’ when we are standing in that bazaar, or at a street corner. Firstly, it requires a different way of looking, and secondly, it requires you to show up. The glimpse, as a visualisation or premonition, gives us a sense of where you might investigate, but you still have to go there.
We cannot tame complexity; only wallow in it.
The character of Baltimore homicide detective Bunk Moreland in ‘The Wire’ described this reckoning with complexity as having “soft eyes”, the ability to squint to be able see the big picture, to not get lost in the detail (though his partner, Detective Kima Greggs, remained unimpressed: “Oh … Zen shit.”)
“If you got soft eyes, you can see the whole thing. If you got hard eyes, you staring at the same tree missing the forest.” (Bunk Moreland in The Wire, S04 E04)
Bunk’s soft eyes technique works for the diagrammatic glimpses in London Squared, given their composition as visual kit of parts at the scale of a city. Yet the tiles also suggest a Powers-of-Ten-like possibility of zooming into the graph. Again, a reduction, but a glimpse that is the start of something.
Other forms of drawing have often got too close to the reduction. Three decades on from the first release of SimCity, architecture schools are still rife with drawings of candy-coloured isometric tiles of various urban components, projected at 30°, to be arranged, aligned and conjoined into functional chunks of model city.
(Admittedly, this includes my own classes at the UCL Bartlett School of Architecture, such as the ‘Incomplete City’ studio, run with Joseph Grima and Marco Ferrari, or more recently with Bryan Boyer at University of Michigan. With ‘Incomplete City’, though, we did at least force students to put their computers aside and co-produce cut-up and taped-together photocopies of hand-drawn elements to introduce variation, subjectivity, chance, and lossy, gritty processes, and ongoing iterative collective drawing, all to avoid the crisp gooey plasticity of rendered CGI and its allusions of perfect knowledge of performance and simple, snapped-together accumulation of complementary features.)
This also highlights the limitation of the screen itself. When we consider the digital twin’s likely realisation at those Forbe’s “product technician’s fingertips” — on a cheap laptop — compared to the other possible formats, we again see how constrained and limited the Twin is. Although hardly comparing Apples with apples, Qing dynasty-era scroll paintings are a richly evocative visualisation of city life. Their scale and form, at 60m long, and containing delicate jiehua (architectural painting) techniques as well as numerous vignettes, enables a constant dizzying shift of scale from detail to landscape simultaneously, in a way no screen can. They enable the glimpse, either of whole or of discrete element, at palace-size scale or via crash zooms into near-microscopic hand-painted detail of individually painted leaves or the ten-thousand ‘long-life’ letters (shou) on a vase.
We needn’t stay in thrall to 300 year-old formats, though. Mobile augmented reality — for instance, in Ericsson’s work with mixed reality for urban planning, with UN HABITAT and Minecraft developer Mojang — offers are a far more interesting mode than most Twins thus far, again predicated on low-res glimpse of Minecraft, as opposed to some false promise of depicting reality. Importantly, using the cellphone as the lens means that augmented reality visualisation happens in the street, in context, and with others.
The kind of performative data that London Squared captures could combine with these augmented reality landscapes to great effect, with its primary user citizens themselves, just as urban data can be generated and acquired entirely on citizens’ terms, as shown by the EC’s DECODE project. (Disclosure: My various teams, at Arup and Future Cities Catapult respectively, worked on early stages of both the the mixed reality and London Squared projects, and I was on the advisory board of the EC DECODE project. Equally, earlier teaching at the University of Sydney offered another variant of drawing data onto the street.)
But perhaps the most important aspect is that this is data that can be used in the street, in a deliberative and social context, immersed with the everyday-complex reality of the street. It ensures that we see that a model is just a model, whilst still making use of data’s ability to pepper a discussion with insights. In Object-Oriented Ontology, Graham Harman suggests that abstract models and data representation will not deal with an era suffused in fake news; only truly engaging in everyday-complex will here:
“Charlatans in politics and elsewhere are best countered not with claims to a truth that no one actually has, but with an unceasing demand that they face up to reality.” (Harman, 2018)
The architectural historian Robin Evans wrote that drawing lies “along the main thoroughfare between ideas and things”, and all these more abstract modes of visualisation usefully hover in-between Morton’s idea that there is data, and interpretation of data. They all provide a glimpse, but all imply work to be done. The tiles in London Squared, with their flat projection loosely mapped, avoid the isometric projection’s dalliance with pretend spatial composition, with faux urban design. They do not suggest that they can be copy-pasted to comprise a community, or to solve the problems of a city like Forrester’s 150 equations. But they might tell us where to show up.
The glimpse, and the more detailed models, readings and interactions that they lead to, is a fundamentally useful tool. The ‘rational units’ of city in this book are artfully visualised at just the right level of conceptual altitude to be recognisable, borrowing the saucy wiggle of the River Thames burned into our brains from 6000 episodes of Eastenders, yet also clearly unrepresentative.
Each unit is a glimpse, and so they are hugely promising as a set of tokens to discuss and develop, comprising a generously open visual toolkit for telling stories about cities. They sit alongside other equally valid representational views — but showing up in the street is where they might be most useful.
The latter is where cities like London actually happen, where those multidisciplinary municipal teams most urgently need to be present, and where visualisations of the past and future can be most usefully generative.
And so here is also where toolkits such as London Squared have most value; making digital twins legible via glimpses, pinned up in a local office window or pawed at on a cellphone screen outside, providing hooks for multiple conversations and interactions, and dropping data into the everyday-complex context of the street, errands and epiphanies and all.
Ed. This is a longer version of the foreword written for After the Flood’s book ‘Cities Squared: Making Urban Data Legible’, which extends the work beyond the London Squared toolkit project I wrote about previously.
Leave a Reply