Tom MacWright

tom@macwright.com

I read A City is Not a Computer by Shannon Mattern on

Review

From the start, full disclosure: I’m in this book, sort of. In its last few pages, A City Is Not a Computer discusses, negatively, Tree Equity Score by American Forests, an application that I worked on. So take what I say here with that grain of salt.

I didn’t like this book very much, for several reasons, even though I agree with a lot of its direction. Dashboards, the idea of data-driven decision making, civic-tech disruption, Sidewalk Labs’s disastrous Toronto project - all of these are, in the general sense, bad. But they’re also all sort of dead. No need to go back to the battlefield.

This book is ostensibly about how abstraction negates nuance, how aggregating statistics and seeing the thousand-foot view of your city is no way to rule, understand, or explore it. That’s pretty correct and reasonable.

But what it comes down to is control of the ways in which people understand the world. Who people listen to, where their worldview emanates.

It could be through the data-driven dashboards, the Transit Screen in their apartment building or the traffic maps or COVID case trackers. It could be through popular culture narratives - an understanding of crime as phenomena by watching CNN, or reading Breitbart, or listening to the radio. You could also understand your city through local word of mouth, through neighborhood friends and family and rumors. Or you could understand it through academic writing, through interpretive dance, through art installations that examine these ideas.

All of these are good, but, here it is, straight-out: the last example, the interpretive art, the academic literature, the nuanced slow cultural way, has been losing for years, if it ever was dominant. The standing of traditional public intellectuals continues to decline. Public art is just as insular, elite, and rare as it has ever been.

Don’t get me wrong: I love nuance. I read a lot, I think, right? I went to college and learned a lot and recommend it to everyone. I love art, books, music, and so on. I’ve been to those spaces, to Living Room Light Exchange in San Francisco and art salons in DC.

But these experiences and materials are luxury goods, right? I have a lot of friends who share a lot of my background – growing up upper-middle class, going to a good college, getting good jobs – who don’t have this media diet, because they don’t have the social connections to the scene or time to read or something else. The cultural-artistic-critic frame is great, but how many people really see through it?

Thinking that somehow indie artistic academic interpretations will become a dominant source of people’s worldview, given how insular they are now, and how books like this one are intentionally written in a particular academic dialect (rampant pluralization of mass nouns, constant citations, lyrical interludes) - how does this work? If you’re trying to convince people of the value of a thing, like trees, like I tried to do, what’s the game plan?

Mattern’s analysis also sets an impossible standard for dashboards. Under her estimation, all dashboard simplify and make data too precise. For example, the dashboard I worked on - the American Forests one - does derive a summary statistic, but breaks out data into 7 different layers. And then more analytics beyond that. And a layer of health data. And satellite imagery. But through her analysis, only the summary statistic matters, and it somehow places the whole project into an uncritical box.

Which is partly a trick based on expectations and assumptions: an underlying assumption is that the consumer of a digital medium like a dashboard is uncritical, whereas someone consuming narrative or art or another format is in a critical mindset. This reflects a sort of nihilism not about technology or the mediums themselves, but about people and their posture towards the world, towards data. That people can’t generate different levels of trust in different dashboards or statistics, or dig deeper for better understanding.

It wasn’t long after building that dashboard that someone mentioned it (not knowing I had worked on it) and said that it was pretty weird that it estimated local O₂ benefits of tree planting. That’s a good and important question - one that I think people should ask, because that estimation is inherently shaky and difficult, and, thankfully, one that people actually do ask. Repeat for crime statistics in the last few years, nearly every conversation I’ve had about the reportedly high crime rates included a lot of discussion about the provenance, texture, details, biases in the data.

And needless to say that plenty of journalists and data scientists actively and enthusiastically interrogate these ideas of uncertainty - see the animated uncertainty graphs in Dark Sky’s analysis or the rich annotations in the NYTimes COVID charts that identify where data is missing or uncertain. Or the many dashboards that make a real effort to be able to drill down, even to the level of an individual person.

Dashboards can mislead and oversimplify, just like narratives can. For example, remember The Omnivore’s Dilemma, a book that layers a rich personal narrative and culture context onto a set of facts, assumptions, and plans that are basically wrong in many ways? Where is the media theory lens that analyses these forms as part of family, rather than marking the digital and the data as an aberration?


The word I’ve been looking for is anti-positivism, which I first encountered in Bike Lanes Are White Lanes. Bike lanes used its anti-positivist bent to pursue the idea that bike lanes were a coded-white symbol of wealth and gentrification, a popular idea in some circles with no grounding in reality. This term isn’t mentioned, but I suspect that it’s the underlying philosophy. Antipositivism (or non-positivism) is a reaction to positivism, in which sociology is practiced through the lens of the scientific method. It swings far, far, infinitely far in the other direction, positing that many things aren’t knowable and quantitative measures are so fraught that they should be avoided or rarely used.

My personal viewpoint is probably slightly preferential to the quantitiative for my own knowledge and investigation, but some balance of the two. But this anti-quantitative backlash to runaway materialist philosophy feels misguided and unproductive, like it doesn’t accomplish any real-world ends and also villainizes its complement.

That’s enough writing for a book review. Let me leave you with a quote, misattributed to Yogi Berra.

In theory there is no difference between theory and practice; in practice there is.

Details