A city fort from above
Fortified City of Ranthambhor ca. 1810–18, https://www.metmuseum.org/art/collection/search/38043

A City Is a City — Against the metaphorization of data

I was invited to give comment (along with Ever Bussey, Jasmine McNealy, and Trevor Owens) on Shannon Mattern’s new book, A City Is Not a Computer: Other Urban Intelligences, at the book launch on 9/15/21. Here is a transcript of those comments (lightly edited and a few hyperlinks added for readability).

(9/15/21 2:31 P.M. MDT)

So… I’m an artist, possibly (?) the only one of us who’d self-describe as an artist on this panel of librarians, writers, teachers, and journalists. As such — and perhaps by definition — I deal with metaphorization. In my own work, metaphors are used as a poetic impulse that sets on things like the city, the dataset, and the computer. Though I use generally use metaphor with the intent to draw useful parallels, in practice they work as often to obfuscate or to complicate ideas even further.

In reading this book (which follows so directly many of my own research interests, from tree-grafting to code licensing) I kept returning to the title, a title that so adamantly rejects metaphor.

A city is not a computer.

The city is also not a tree, a dashboard, a business, or a luxury good. We are temped briefly by an organism but are reminded that despite its waste-streams and circulatory system a city is not a body. It is not a machine made of replaceable discrete parts, or a factory made of reparable discrete machines. The city is not a brain and the city is not a palimpsest and the city is not the internet and the city is simply not a computer.

Stepping back, each of these metaphors serve (in a undergrad poetry class kind of way) but are also absurd; of course a city isn’t a tree, or a business, or a body, or whatever. Get your head out of the Ferdinand de Saussure, Everest.

But metaphor isn’t empty (I should know this) and the city as luxury good or factory or dashboard has an agenda. It grants power and agency to a particular kind of manager, mayor, tech lab, real estate developer, military general, or financial investor. It says — ah, this city is like this — and I know just the person to fix it, grow it, watch it, maintain it, run it, mulch it, or whatever other verb fits the particular metaphor of the moment.

A whole chapter of A City Is Not a Computer is dedicated to the city “dashboard” — a visualization metric of data that has certainly had its moment this year in the myriad of city and state pandemic dashboards, which I still check religiously before heading into town for my weekly grocery trip. In the book, Shannon traces the history of “the dashboard”, from the literal running board of a horse-drawn carriage (meant to keep the mud out the dashing hooves) to car dashboards with their limited set of fuel indicators and lights to the fighter planes of both World Wars and eventually to the city terminal, a now-digital interface full of dials and charts labeled things like “investment flow” and “happiness index”. Made to advertise a panoptic, sci-fi vision of a spage-age mayor’s office, the contemporary smart city dashboard is as much for show as it is functional. Sure, the mayor can see there is congestion on the 405 from their office — but what, exactly are they supposed to do about it?

These days, a data dashboard actually intended for a human to take fast-moving policy action from is almost a quaint idea. This is because so many of these systems are automated. In an era of microtransaction bots housed in datacenters close to the New York Stock Exchange, the stock trades well before the Bloomberg Terminal displays the transaction. This is true of automated traffic lights and “hotspotting” crime predictive systems and currency exchanges and the calculations that go into your credit score. It all happens faster than human time. A dashboard is no longer a control panel — it is a newspaper. Forget if what it displays is the whole picture — it isn’t ever, anyway — when a dashboard operates as a report after the act with little capacity to interject, does it matter if what it displays is even true?

The dashboard’s relationship to the fighter plane is relevant here, as is the smart city’s insistence on the “utopian” idea of supplanting the manual labor of things like, say, coffee delivery to the drone. Both have a direct relationship to violence, and automation that acts and then reports the action is violence personified. Firing an explosive shell was once as distanced from the lived reality of murder as pressing a button on a dashboard. Now it is as far removed as seeing that the shell has struck its target, and, later, finding that the target was an aid worker carrying nothing but water — not, as the autonomous computer vision system claimed, bombs.

The “data” here was an excuse to produce harm, as is so often the case. This is also true in the broader sense of metaphorizing the city as a dataset, something Shannon has been very clear to highlight throughout the book. There is an express violence that hot-spotting, crime prediction, redlining, and the often no-less-violent actions of “urban renewal projects” allow the city to commit while under the auspices of “…but the data says!”

We’re reminded that under no conditions will the city be called a dataset.

Rather than attempt to widen the lens of city dashboards and health index pie charts to reflect the lived experiences of the broad swaths of humanity they attempt — and fail — to accurately represent, the time would be better spent internalizing that these tools have a purpose and it is rarely if ever to actually reflect those lived experiences. Personal data from marginalized populations does not lend itself to data visualization not only because it is systemically under- or miscounted but also because data visualization is — or directly borrows its language and tooling from — white supremacist projects, including those of weaponry, industry, mapping, demography, and even the colonial project of the highway system.

There is no dataset that can fundamentally subvert this.

This is because data will be skewed by power wherever it exists; at least under capitalism, empire, and the carceral state. The very practice of “collecting data” (empirical data, what is envisioned as “true” data) is a project of a worldview that believes in a single empirical truth. The city as a dataset — even when operated with best intentions by what I can only assume are very overworked civic servants — intends to subsume the individual experience into a pie chart that can be read, at a glance, by someone in power.

Data cleaning — a practice that I have spent hundreds of thousands of hours on in my own creative work — has an agenda. When you throw out what is “dirty” about your data, you are keeping what is clean; a troubled terminology if I’ve ever heard one. Even in my artistic practice, this comes at cost. I am calling something useful, poetic, interesting, or fundamentally more than the next thing over. Now, I don’t hand out bank loans. But I do contribute to the project of cultural memory. And if you scale this practice to the size of a city you can understand why we are in trouble.

This is smart city thinking. This worldview looks at city services — transit, garbage, water, public works, social services, health — and wonders if the city could be more efficient and better organized if these systems were “smart”. It thinks that maybe if these systems tracked themselves, understood how it all fit together, saw wasted hours and duplicated routes put up on a big screen in an office somewhere, perhaps all with a friendly AI directing traffic and reordering supplies and probably making up the weekly employee schedules that always somehow leaves everyone exactly 1 hour short of-full-time, then maybe the city would achieve true efficiency.

But these systems already are smart. They are all operated by people with intrinsic knowledge — and not just where that one neighbor hides the recycling cans, but also the names of kids and the birthdays of dogs and the needs of elders who can’t make out to the store anymore. This knowledge — human, caring knowledge — does not make it into the “smart city dataset”. But I promise you, give sanitation workers a raise and an extra day off every week and truck that starts up reliably in the morning and less routes to do in more time, the city would be better — if not more efficient.

The smart city always attempts to build from scratch (from the fantasy of an empty platform, a used industrial space, a barren lot) but there is no scratch. People lived there. If they’ve been forgotten, it was by design. Animals and plants lived there. If they’ve been rousted, it was with poison and traps. The rats and moths and feral cats that undoubtedly still live there have a knowledge. The smart city doesn’t see them, and if it does, it doesn’t care except where they are cost or benefit.

And in truth, this is how the smart city sees humans too.

Perhaps out of all of these myriad of metaphors, a city as an ecosystem comes closest — an ecosystem will sprout unbidden just about anywhere that things live, which is to say — everywhere. You don’t have to think long about the particular cultures of subway bacteria or the local population of pigeons before the city becomes like a forest. But even this metaphor — as broad as it is — has its own agenda.

A city as ecosystem greenwashes that smart city project, buries the surveillant aspect of a techno-organic system of control in language about care and ecology. The smart city embeds sensors in the smart soil to tell the sprinklers when to turn on. This saves water, perhaps, but also builds a wireless-connected network of discrete powered devices constantly pinging home with hyper-specific data about a particular patch of earth; data that can be repackaged and sold, per usual, to the highest bidder.

I’m reminded, as I am practically every week, of that deeply cursed blockchain experiment about giving animals money. (The solution, and I cannot stress this enough, to trees and animals being ignored by our current wealth-driven surveillance tech is not to find a way to monetize the trees and animals and thus surveil them too.)

Capitalism can eat just about anything, and the fantasy of the smart city is damn well trying. Because when knowledge enters the formal, cleaned spreadsheet of the dataset it becomes again an asset. This extractivist approach to dealing with climate horror should be familiar by now; a common refrain spouted from smart city projects and their ilk is something along the lines of “we study the ignored knowledge of marginalized and indigenous communities to incorporate it into our shared future”. But this is done generally without offering those same communities any power in shaping said future. Historical knowledge turned asset in the accumulation of data.

The greenwashing of the smart city is so effective partly because of the role that data does play in environmental lobbying. Things like annual migrating bird kill counts and sea ice measurements are among the only tools ecologists have to deploy against a system set deeply in favor of capitalist gain.

To the eyes of U.S. law, there is never a rare salamander in the path of the oil pipeline unless someone reputable has published a dissertation describing exactly this creature ; and the oil pipeline shall not be halted for anything less than a violation of the Endangered Species Act (whatever else is in the way — common newts, beautiful meadows, sacred land be damned).

When the onus of proof of value rests on the salamander and not the pipeline, the salamander is entered into the dataset to become a bureaucratically readable, empirical truth that can (every once in a while) divert a pipeline a few miles south. These few hard-earned condolences to “ecology” work to coat the miasma of the smart city in this kind of rhetoric or thinking; that if only we had more data about what trees needed water when — or even if we made them assets, to be traded at market — perhaps then they could thrive in this hell being built around us.

But this is absurd. Rather than technologizing the smart soil of urban gardens we can simply collect the rainwater off nearby roofs and water them when the leaves begin to droop (plants come bundled with their own sensors and displays — ones that tell far more, with more directness and clarity than any dashboard).

Rather than calculating the street-tree and shade relationship to the poverty index of Los Angeles neighborhoods in order to “instigate change on the ground”, software think-tanks and the cities that employ them can simply provide residents access to seedlings (and the resources to keep them alive) if these trees are wanted. Or even better- they could ask what would be helpful in a case by case basis because, again, humans come equipped with language which can explain far more in far greater detail than an aerial study of tre canopy and home value.

Or — to make it even clearer — in the case of climate change, we truly have all the data we need. The reports are clear. It is calamitous. You don’t even need a dataset to see it — I can walk up the hill from my home and see the wildfire smoke. I can feel the heat of the summers and the force of the hurricanes. Yet just this week the Biden administration claimed that the devastating IPCC report “does not present sufficient cause” to pause an expansion to offshore drilling.

This is not due to a lack of data. The data is there. It is instead because there is no empirical, data-driven value that you can assign to things like “a life of dignity and abundance in a culture that wishes to see all people thrive” that squares it against capital gain. There is no empirical, data-driven value that you can assign to this at all.

The world is dense and complex and layered, broad and swirling and impossible to force into a dataset — or even a metaphor — but our relationship to need both as individuals and as governing systems made of individuals does not have to be complex, and should not be.

You can simply give people money when they are poor. You can simply give a tree water when it is thirsty. You can simply overturn the military capitalistic project of the United States of America before it destroys society. (I joke. This is not so simple. But, also- it kind of is. Land Back is a simple and clear idea.)

You can simply care for what you care for, without ascribing it a value at all.

And that might be the core message of The City Is Not a Computer- that despite all the obfuscatory metaphors and ontologies and datasets and investment projects that graft themselves onto cities — no matter how those are used and abused, abandoned and remade — the project of maintenance is going on all around us as individuals tend to the needs of what they care for.

A City Is Not a Computer reminds us that a city is a city, and a city is — and always will be — simply a place where people live their lives.





Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium


A City’s Chance to Capture Jobs in the 21st Century

Protecting Intellectual Property in the Digital Age and Beyond

Are Robots Ready To Replace Us In Jobs?

Flying Cars aren’t Coming Any Time Soon

The Future of Robots

The Top 5 Technology Trends for 2022: The Year of Decentralisation

The Top 5 Technology Trends for 2022: The Year of Decentralisation

Impact of industrial automation for the manufacturing sector

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Everest Pipkin

Everest Pipkin


More from Medium

Responsible AI at Scale with Beth Rudden

Women in AI Ethics podcast series sponsored by IBM. Building Responsible AI at Scale. Beth Rudden, IBM Distinguished Engineer & Principal Data Scientist — Cognitive & AI Services

Addressing Climate Change Will Require Prioritization by CEOs

Has tech innovation already moved us towards a more sustainable future?

Can we measure (and predict) religious change?

A crowd of people moves through a large, open train station