Kevin Webb Profile picture
21 Feb, 26 tweets, 8 min read
This is a fascinating story that, even putting aside the specifics of what happened in Portland, offers useful insight into the state of side-streamed, citizen generated data in public policy (tl;dr things are not in a good way)...…
I suspect this story is part of a broader shift: we're reaching the end of the line for surreptitiously collected, citizen generated data as a policy and regulatory input.

That idea, at the core of the whole "big data"/"smart cities" edifice, is not working.
First off, uncalibrated data isn't just not useful in a policy context, it's actively harmful. Particularly in places where we're dealing with structural inequities.

In this regard I'm actually a big fan of Replica's work, as it began with tackling the calibration issue upfront
But calibration, as discussed in this story, is a public responsibility.

User data collected in a consumer application is not a census. It is, at best, a potentially useful sample *of the telemetry* generated by the users *of the specific application* that collected the data.
This may come as a surprise because pitches from vendors for the last 20+ years have been telling us "everything has changed"

A decade ago @mtchl eloquently summarized the fallacy of this kind of pitch, yet, it's just as seductive today as it was then…
Vendors have been telling us "big data" means we're now able to measure ~100% of the population in real-time with sufficient precision that the last few hundred years of lessons in statistics and quantitative social science no longer apply. That was, is, and always will be a lie.
The attraction of this pitch is that it's it leads us to believe that understanding human behavior is no longer a messy process of inquiry, but rather that human behavior is an objective phenomena that can be measured and managed.
This pitch isn't new. It's not even the first time it's been applied to urban policy. It was at the core of our past attempts at high modernism in city building.
It should come as no surprise that today's fantasy of urban "digital twins" finds its origin in industrial control systems and architecture, just as did past attempts at deploying new technologies and methods to creating orderly cities.…
This isn't to say that measurement isn't useful. It's just hard, and begins and ends with the reality that there's no single objective measurement for complex systems.

The application of quantitative methods to cities is a way of thinking, not a number
And this is where things have gone wrong: public investment in ways of thinking about challenging policy questions have reached near crisis levels. And vendors are now selling into that void with tools that offer the appearance of quick solution to decades of underinvestment.
But rather than collecting new uncalibrated sources of user generated data we should focus on expanding our methods of calibration. This is expensive, and starts with refinement the questions being asked (there is no such thing as universal calibration).
It also cannot be a black box. The way we arrive at the number is substantially more interesting and important than the number itself. The methods of analysis are inherently part of the process of policymaking and must be open to scrutiny and debate
This leads to the second challenge: the extraordinary privacy issues related to citizen generated data, especially the kind relevant to transport policy, are at odds with the level of inquiry needed to calibrate for policy-relevant applications. The data can't be shared.
In fact, this data shouldn't be collected in the first place.

Fortunately, as the public (and legislators) become more aware of the collection methods many of these practices are on the verge of becoming illegal.
Using organic citizen-generated data in a policy context is potentially useful, but it depends on a level of governance (and self-restraint) that we appear to be incapable of in the US. And it cannot be built on the current, surreptitious and abusive collection methods.
But rather than placing smart limits collection and use, the US urban policy community hitched their desire for unrestricted access to citizen-generated data to the argument that abusive corporate-customer relationships indicate that citizens right to privacy no longer exists.
That argument was not only incredibly cynical, it was poorly informed and un-strategic given that many policy applications depended on precarious and abusive data laundering schemes that existed only because public didn't know about them.
As that awareness has grown both legislative and industry self-regulation are bringing these practices to an end, and will take many of the businesses and analysis methods that depended them down with them.…
On one hand this will eliminate existing data sources (if your city buys citizen-generated traffic data, do you know where it actually comes from?), but on the other it will force a more explicit and necessary conversation about how and why we measure urban systems.
In addition to tackling questions of provenance and calibration, one of the consequences of this will be disentangling insight from enforcement applications. Confusion between these two as at the core of some of the more pernicious privacy challenges in making use of citizen data
We can't fix the privacy issues with using data for insight because the actual goal is often enforcement (sometimes euphemistically referred to as "optimization" urban policy circles), which is inherently invasive, if not outright illegal.
I gave a talk last fall for the OECD-ITF about this confusion, and its consequences and remedies. Breaking these goals apart is the first step to finding a solution.…
In the meantime, due to this confusion, the US urban policy community is currently barreling towards losses in court that will not only limit current and future use of citizen generated data in policy, but likely undermine necessary forms of corporate oversight.
As we pick up the pieces from those losses, I'm hopeful we'll build a more intentional and intelligent public data collection infrastructure, that rather than side-streaming dangerous and inadequate consumer data, thoughtfully measures the things we actually want to know.
As a final note, I highly recommend following @KateKayeReports. She's one of the best reporters on the urban tech and surveillance beat today. In addition to her coverage of Replica, her podcast series for @smartcitiesdive is fantastic:…

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Kevin Webb

Kevin Webb Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @kvnweb

19 Feb
“With the exception of straight-on climate change issues, in general in smart cities, we’re not actually tackling the big issues. And the big issues have all been thrust out into the open now through the pandemic. Racism, inequality – that sort of thing.”
I appreciate this piece and agree that there's a (needed) recalibration on "smart cities" as tech solutionism for, or tech deflectionism(?) from the big issues.

But the smart cities conversation is still too focused on tech infrastructure as a buyable thing...
...the kinds of infrastructure we should be talking more about about are emergent, and transform/intermediate the ways we experience our communities. While a lot of the smart cities pitch is about buying tech to give the illusion of control, and at best, a return of status quo...
Read 7 tweets
26 Aug 20
To add to the week's Twitter conversation, I to follow up on the IRL conversations I've had today with folks concerned about things becoming too personal, and @jfh's concern that the things I'm saying are distracting from a more needed conversation about "big ideas." Agreed 100%
So in the interest of explaining how things got to where they are and why we're apparently now unable to talk about big ideas, I'd like to share a recent email exchange with Jascha's colleague at Lacuna, Sean:
The memo I referred to and, that Sean acknowledged helping write (and now regrets), arrived just a short while after this op-ed was published in summer 2019…
Read 37 tweets
9 Jan 20
I've been fascinated private covenants. Over the summer I bought a house with a racial covenant in the deed and was shocked by how successful the authors had been in preserving segregationist land use policy even after the language tied to race was rendered inert.
The racial covenant (item g.) was struck down in Shelley v. Kraemer two years after my deed was drafted but items a-f remain in effect today and are still celebrated as foundations of modern of land use policy (ADU limits, setback requirements, and construction cost floors)
The language for all of these provisions were written by the same person with the same goal. It was a diabolically crafted plan to keep neighborhoods segregated long after the racial covenant was struct down. And it worked.
Read 9 tweets
21 Feb 19
Seeing a distinct "data populism" spike over the last year. On an emotional level it's an understandable reaction to profound inequality that's emerged w/ platform monopolies but early-stage responses (LA's "active management" of citizens, CA's "data dividend") are really bad. /1
Top line takeaway: progressive policy folks (save the @openmarket crew) lack a coherent understanding both the tech itself and contemporary vertically integrated firms. It's unclear how anyone plans to build responses to processes they fundamentally don't understand. /2
Pretty much everyone that gets how this stuff works is now employed by one of the current/hopeful platform monopolists. Public minded people are taking these jobs as a defensive move because of institutional/political resistance to understanding broader economic/power shifts. 3/
Read 24 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!