It's like KittenWar for urban spaces

September 28, 2011 1:51 PM

Which place looks safer? Which place looks more unique? Which place looks more upper-class? MIT is crowdsourcing a "perception network" to analyze people's subconscious judgments about urban spaces. Preliminary results for Boston, New York City, Vienna, Salzburg, and Linz (Austria).

Visual preference surveys are a common tool used by urban planners to gauge public interest in and desires for the aesthetic components of their neighborhood. This PDF is a good example of a summary report and shows the most/least desirable images from a study in Bellevue, Washington.

This article from the Journal of the American Planning Association (PDF) goes into detail about how hierarchical models are used with visual surveys to improve the streetscape of an area.

posted by desjardins (45 comments total)

People like trees.
posted by smackfu at 1:56 PM on September 28, 2011 [11 favorites]

Also: the Austrian photos have a better picture quality while the American ones look washed out and JPEG artifact-y.

Real-life datapoint: I lived in Salzburg for a year and never felt unsafe. I can't say that about American cities.
posted by dunkadunc at 1:58 PM on September 28, 2011 [1 favorite]

My guess: "Safer" is so strongly correlated with "front lawns and/or blue skies" that everything else will be dwarfed.
posted by Tomorrowful at 2:03 PM on September 28, 2011 [2 favorites]

People like windows, too. Startling how much that seems to matter.

Would be especially interesting to see how an answerer's demographics affected his responses. Do all people find wide open spaces and uninterrupted walls threatening, or is that culturally-influenced?
posted by foursentences at 2:03 PM on September 28, 2011 [1 favorite]

I had a choice between a free way tunnel and a hospital, so the choice seemed obvious. Although I didn't consider the high risk of infection!
posted by Think_Long at 2:04 PM on September 28, 2011

Also: the Austrian photos have a better picture quality while the American ones look washed out and JPEG artifact-y.

Real-life datapoint: I lived in Salzburg for a year and never felt unsafe. I can't say that about American cities.

Well of course you didn't feel safe, not with all those JPEG artifacts loose on the street.
posted by Bulgaroktonos at 2:04 PM on September 28, 2011 [10 favorites]

It asks for your age and gender if you do a few in a row. I think also pictures with people in them would look safer.
posted by SpaceWarp13 at 2:05 PM on September 28, 2011 [1 favorite]

With lawns, you always know some is watching to make sure you stay off them.
posted by Kabanos at 2:05 PM on September 28, 2011

Q: Which place looks safer?

Option One: Some nice apartments with hedges
Opion Two: Standing in the middle of a highway

Gosh, I wonder.
posted by Winnemac at 2:12 PM on September 28, 2011 [5 favorites]

Regardless of the weather in each photo, I was looking for sidewalks and residences, as opposed to creepy alleys and expanses of pavement with few buildings around. I've seen that horror film.
posted by polywomp at 2:13 PM on September 28, 2011 [1 favorite]

Oddly enough, I was unable to rate more than a few when asked which looks more unique, or which looks more upper class. The pictures look the same to me in that context.
posted by polywomp at 2:15 PM on September 28, 2011 [1 favorite]

Q: Which place looks safer?

Option One: Some nice apartments with hedges
Opion Two: Standing in the middle of a highway

Gosh, I wonder.


Seriously. Originally I was excited by the prospect of what MIT would do. The actuality? Badly biased samples.

I guess it only takes a little more imagination to realize that sampling peoples' subjective response to a minute spectrum of any city's urban realm is just pointless.
posted by Reasonably Everything Happens at 2:19 PM on September 28, 2011 [1 favorite]

Is it badly biased samples, or just that we are actually have very strong opinions on what looks safe or not?
posted by smackfu at 2:23 PM on September 28, 2011 [2 favorites]

I realized pretty quickly that my first impulse was to label whichever one had better lighting as safer. After that, my conscious brain kicked in and I started thinking about how many eyes would be on the street and whatnot.
posted by craichead at 2:28 PM on September 28, 2011

This seems like a ridiculously bad research design.
But how could it be done? The questions are really complex and hard to separate in a manner acceptable for research. I think some of the relevant questions are already answered in Life Between Buildings, and in Jane Jacobs' work. But there is still work to be done. Maybe it would be a good idea to start with a (respectful) overview and critique of the existing research, so we could like, start on a higher platform.
posted by mumimor at 2:32 PM on September 28, 2011 [1 favorite]

I've facilitated one of these surveys, and of course there is bias. The thing is, if you ask someone whether they would rather live in mid to high density mixed-use areas with narrow streets and traffic calming and crowds of people everywhere, or would they prefer traditional suburban environs with plenty of turning lanes and huge setbacks, many will in fact choose the latter.

But when you show folks what the former can actually look like, and what the latter often looks like, as opposed to how they may perceive them, that's often a whole other story. Or sometimes it isn't. In my experience it doesn't work well in a vacuum but it works well as a dialogue starter.
posted by gordie at 2:43 PM on September 28, 2011 [2 favorites]

Something is either unique or it is not. It is a binary quality. Something can be more or less obviously unique, or more or less remarkably less, but it cannot be more or less unique, just as cannot be more or less infinite. It either is, or it is not. /pedant
posted by oneironaut at 3:02 PM on September 28, 2011 [5 favorites]

"More unique"? I wish I could vote for which photo looked like it understood English.
posted by RogerB at 3:02 PM on September 28, 2011

In one of the "safer" questions my choices were

1. A busy looking street with no markings
2. A quiet looking street with a big zebra cross walk.

I chose 2 but, here's another way you could describe those same pictures

1. Nice looking buildings, fancy-ish cars
2. crumbling facades, chainlink fences, one crappy car.

So, the danger that I was focusing on was, "How likely am I to get killed crossing this street?" rather than, "Am I risking getting mugged or propositioned if I stand on this corner?"

So, my critique of this study is, how has MIT operationalized "safe?"
posted by bilabial at 3:04 PM on September 28, 2011 [2 favorites]

oneironaut: "... just as [it] cannot be more or less infinite. It either is, or it is not. /pedant"

Actually, there are degrees of ininity. There are e.g. infinitely more real numbers than there are natural numbers, and there are already infinitely many of those. /mathpedant
posted by brokkr at 3:09 PM on September 28, 2011

there are degrees of infinity

that's daft
posted by the cuban at 3:15 PM on September 28, 2011

Is America like 90% concrete? Didn't you guys have a forest at some stage, the one where that bear with the hat lives?
posted by tumid dahlia at 3:18 PM on September 28, 2011 [3 favorites]

Is America like 90% concrete?

That's sort of like reading your question and asking whether Australians are 90% dingbats. These photos aren't exactly a representative sample of American spaces.
posted by craichead at 3:31 PM on September 28, 2011

there are degrees of infinity

that's daft

Nope, it's a fact. There are things which are countably infinite (the natural numbers, rational numbers) and things which are uncountably infinite, (the real numbers).
posted by empath at 3:31 PM on September 28, 2011

that's daft

I take it you've never heard of Cardinal Numbers? There are an infinite number of integers, and there are also an infinite number of real numbers between each integer (and between every pair of real numbers)
posted by delmoi at 3:32 PM on September 28, 2011 [1 favorite]

We need all that concrete. Its not like we get to ride kangaroos to work.
posted by Kloryne at 3:35 PM on September 28, 2011

I'm pretty much just going to go on believing that it's 90% concrete. You don't get to cherry-pick your photos any more, Yanks! "Oh look we've got this nicely-lit statue, it sprang up out of the ground because of all the liberty." Ha ha, we know what's really going on!
posted by tumid dahlia at 3:38 PM on September 28, 2011 [3 favorites]

I mean, here's me sat at home thinking Boston was that nice snowy English-looking place from Fringe, but it's just a big car park, really, isn't it?
posted by tumid dahlia at 3:39 PM on September 28, 2011 [1 favorite]

On:
Option One: Some nice apartments with hedges
Opion Two: Standing in the middle of a highway

It might be a sort of sanity check, when opening your system inputs to the public you do want something to filter out nonsense responses.

Or it could be the yet another result of the Frogger players of American's recent gains in influence and prestige, they've even got MIT touting their nonsense!
posted by oblio_one at 3:50 PM on September 28, 2011 [1 favorite]

Degrees of infinity. Nifty. The unique criticism holds, though.
posted by oneironaut at 4:26 PM on September 28, 2011

Big trees and lots of cars make a place look safer to me. In reality, these things probably just provide places for evil to hide.
posted by limeonaire at 4:47 PM on September 28, 2011

Is America like 90% concrete? Didn't you guys have a forest at some stage, the one where that bear with the hat lives?

We're paving them as fast as we can but it's a big country. Thank you for your patience, you can buy flags in the gift shop on the way out.
posted by qxntpqbbbqxl at 5:25 PM on September 28, 2011 [9 favorites]

Which website looks more low-brow?

Weekly World News | USA Today

Newspapers are the same

posted by obscurator at 5:27 PM on September 28, 2011

I think my safety perception must be a little warped. First I checked to see if it looked like somewhere I might get shot or mugged. Only two of the bunch I clicked through even slightly fit that category, though many I thought you couldn't see enough of the street to judge well.

Then I thought about how likely I was to get hit by a car as a pedestrian. Several fit there.

Those seem like pretty different measures of safety, though. And if neither picture had fit either "unsafe" category, I didn't perceive any difference in safety.
posted by sepviva at 6:16 PM on September 28, 2011

Seriously. Originally I was excited by the prospect of what MIT would do. The actuality? Badly biased samples.
So, my critique of this study is, how has MIT operationalized "safe?"

No, but you see, it's the other way around -- these images are probably randomly taken from Google Street View and processed through the filter of 'the crowd'. 'Safe' and 'Unique' and 'Upper-class' are deliberately ambiguous because they're trying to find out what the definition of these words are, according to you.
posted by suedehead at 6:55 PM on September 28, 2011 [4 favorites]

Big trees and lots of cars make a place look safer to me. In reality, these things probably just provide places for evil to hide.

As a woman who walks pretty much every where I can and home from work at pretty late at night, I find that both fancy corporate business districts that empty out at 5:00 and upscale, leafy neighborhoods without well-lit sidewalks don't feel all that much safer than the parking lot of the sketchy 24 convenience store across the street from the overpass if you have to walk through them at night.

Or put more succinctly: show me these same locations at 11 pm on a weeknight.
posted by thivaia at 7:38 PM on September 28, 2011 [1 favorite]

If the landscaping looks neat and kept up and lushly green, I feel safe. I like more people rather than less.

I also got past the lighting and focused on the areas, and found that wide open spaces that looked both unkept (scraggly or dead or no vegetation, broken fences, etc.) and like no one was around for miles made me uncomfortable.
posted by misha at 7:54 PM on September 28, 2011

Great idea, badly executed. I can't get a feel for some of these places because of the wicked distortion in the photos. I was hoping for some kind of panoramic or even 3D image.
posted by storybored at 8:04 PM on September 28, 2011

Q: Which place looks safer?

Option One: Some nice apartments with hedges
Opion Two: Standing in the middle of a highway

Gosh, I wonder.

Being a pedestrian in many North American cities often does feel like being in the middle of a highway. It's not a sampling issue, it's a reality issue.
posted by parudox at 8:27 PM on September 28, 2011 [1 favorite]

Being a pedestrian in many North American cities often does feel like being in the middle of a highway.

I avoid driving whenever possible, so I'm sympathetic, but we're not talking about Kansas city here. Boston and New York are both quite accessible to pedestrians. The drivers are the ones to pity.

Places like this or this are just weird to vote on. With most of the pictures you are asking "would this be a nice neighborhood to live/work in?". Then one of these guys come along and suddenly the question no longer makes any sense. These aren't even places where you are supposed to be stationary. It messes things up a bit, because in reality tourists would not often consider the highway a scary dangerous area to be avoided, unless they were inexplicably stopped in the middle of it like this.
posted by Winnemac at 11:25 PM on September 28, 2011

Degrees of infinity. Nifty. The unique criticism holds, though.

I hear "more unique" all the time. It's always obvious that what people mean is that some things are more unlike other things than other things.

Words sometimes carry multiple meanings. "Pedant", for example, can often be read as a convenient shorthand for "person on the Internet who is unhappy that words carry multiple meanings".
posted by brennen at 3:00 AM on September 29, 2011

I don't have it on me, but I pretty sure there's a chapter in Kunstler's Geography of Nowhere (or possibly Home from Nowhere) in which he talks about visual preference surveys and notes that though people are almost unanimous in preferring mixed-use, walkable neighborhoods with lots of trees, that's mostly not what we build.

As an urban planning student I facilitated one of these in a community participation meeting. The area in question was along a wide road with a grassy median. There was a mix of high density residential and light industrial uses, all with huge setbacks and a pointless sidewalk. We showed them pictures of various streetscapes from around the country. People loved street furniture, landscaping, and buildings closer to the street. That meeting was 5 years ago, and this is what it looks like today.
posted by desjardins at 7:17 AM on September 29, 2011 [2 favorites]

There are an infinite number of integers, and there are also an infinite number of real numbers between each integer (and between every pair of real numbers)

That's all true, but maybe it can be taken to imply something false, namely that this phenomenon — real numbers between real numbers ad infinitum — is the reason that how-many-real-numbers-there-are is a bigger infinity than how-many-integers-there-are. But on the contrary, there are rational numbers between rational numbers ad infinitum, but how-many-rational-numbers-there-are and how-many-integers-there-are are the same infinity.

I guess I'm contributing to a derail here, but math imposes its own obligations.
posted by stebulus at 8:36 AM on September 29, 2011

US CITY DWELLERS DEFENSIVE ABOUT TRASH-LITTERED STREETS, SHODDY PAINT JOBS AND DEAD LOOKING PLANTS FILM AT 11
posted by DU at 7:21 AM on September 30, 2011

But on the contrary, there are rational numbers between rational numbers ad infinitum, but how-many-rational-numbers-there-are and how-many-integers-there-are are the same infinity.

Yeah, and there are ways to count off all the rational numbers given an infinite amount of time.

The difference between that and real numbers is that no matter how you decide to order the real numbers, there is always a way to construct a real number which your counting method will never reach, even given an infinite amount of time.

(not that you don't know this, just clarifying why rational and real numbers are different)
posted by empath at 10:11 AM on September 30, 2011