CNTRL_F+"Race"=0 Results
February 13, 2015 7:18 AM Subscribe
Citylab on the new data tool PlaceILive: "While PlaceILive is obviously a more serious endeavor than some of those awful "ghetto-tracker" neighborhood apps we've written about before, (previously) the site does wcore areas with high poverty rates and low educational attainment levels lower on its Life Quality Index score.
"We just genuinely believe that if there are more educated people, it is a nicer neighborhood, and the same with income," Legeckas says.
That may not necessarily be true, but it points to a broader problem with statistics-driven quality-of-life measures—one that Legeckas acknowledges: numbers can be deceiving."
"We just genuinely believe that if there are more educated people, it is a nicer neighborhood, and the same with income," Legeckas says.
That may not necessarily be true, but it points to a broader problem with statistics-driven quality-of-life measures—one that Legeckas acknowledges: numbers can be deceiving."
You know, it says that numbers can be deceiving, but it doesn't actually show how. I think a life quality score of 70 for 620 8th Ave is eminently fair.
posted by corb at 7:37 AM on February 13, 2015 [1 favorite]
posted by corb at 7:37 AM on February 13, 2015 [1 favorite]
The whole site apparently only works for Chicago, SF, NYC, London, and Berlin. Their primary markets are therefore either hipsters or oligarchs, which probably skews the analysis somewhat.
posted by Kadin2048 at 7:44 AM on February 13, 2015 [6 favorites]
posted by Kadin2048 at 7:44 AM on February 13, 2015 [6 favorites]
How It Works: We slow your browser to a crawl, and we don't deign your city worthy of our consideration, so you hit the 'back' button irritated.
posted by Mayor Curley at 8:20 AM on February 13, 2015 [8 favorites]
posted by Mayor Curley at 8:20 AM on February 13, 2015 [8 favorites]
The whole site apparently only works for Chicago, SF, NYC, London, and Berlin. Their primary markets are therefore either hipsters or oligarchs, which probably skews the analysis somewhat.
You do realize these are the biggest cities in the world right? I feel you're being critical for the sake of critical, considering these cities are home to millions of people all along the social and economic spectrum.
posted by kurosawa's pal at 8:26 AM on February 13, 2015 [1 favorite]
You do realize these are the biggest cities in the world right? I feel you're being critical for the sake of critical, considering these cities are home to millions of people all along the social and economic spectrum.
posted by kurosawa's pal at 8:26 AM on February 13, 2015 [1 favorite]
Interesting. From the SketchFactor article: "SketchFactor is exclusively focused on improving city exploration on foot".
This is essentially the same mission as a whole genre of dérive apps that try to guide you (probably hipsters, but still) into exploring a city in a much less prejudicial way. Hopefully a way that doesn't lend itself to this kind of redlining. In part, this seems like a result of how the app does the guiding: SketchFactor and PlaceILive both use statistics to decide what neighborhoods are nice; Dérive app uses a randomized "deck of cards".
I wonder if the prejudice is already there in the statistics, like in how they're collected, or a natural part of how we look at the statistics. Either way, the statistics are about things "have been" in these cities, not so much about how they "could be" or how we could change the places we live.
posted by wormwood23 at 8:38 AM on February 13, 2015 [1 favorite]
This is essentially the same mission as a whole genre of dérive apps that try to guide you (probably hipsters, but still) into exploring a city in a much less prejudicial way. Hopefully a way that doesn't lend itself to this kind of redlining. In part, this seems like a result of how the app does the guiding: SketchFactor and PlaceILive both use statistics to decide what neighborhoods are nice; Dérive app uses a randomized "deck of cards".
I wonder if the prejudice is already there in the statistics, like in how they're collected, or a natural part of how we look at the statistics. Either way, the statistics are about things "have been" in these cities, not so much about how they "could be" or how we could change the places we live.
posted by wormwood23 at 8:38 AM on February 13, 2015 [1 favorite]
"You do realize these are the biggest cities in the world right? I feel you're being critical for the sake of critical, considering these cities are home to millions of people all along the social and economic spectrum."
*cough*Tokyo Seoul DF Beijing Delhi Cairo LA*cough*
posted by klangklangston at 8:42 AM on February 13, 2015 [6 favorites]
*cough*Tokyo Seoul DF Beijing Delhi Cairo LA*cough*
posted by klangklangston at 8:42 AM on February 13, 2015 [6 favorites]
Mostly what I notice is that the PlaceILive web site isn't nearly granular enough for Manhattan. Assigning the same "life quality index" score to, say, 104th Street at both West End and Manhattan Avenues doesn't reflect the fact that these two locations, despite being separated by only four blocks, have vastly different residency experiences on almost all metrics.
posted by slkinsey at 8:57 AM on February 13, 2015 [4 favorites]
posted by slkinsey at 8:57 AM on February 13, 2015 [4 favorites]
I wonder if the prejudice is already there in the statistics, like in how they're collected, or a natural part of how we look at the statistics.
More than either of those, I think it's just that the prejudice is baked into in society. Even assuming there are problems with how these factors are calculated or how they're presented, there's nothing wrong in principle with people desiring neighborhoods with low crime rates and more highly-educated residents. Still, because those variables correlate somewhat strongly with ethnicity (due to racial minorities being targeted by law enforcement and denied opportunities for many generations), there is functionally no difference between desiring neighborhoods with these characteristics and desiring neighborhoods with very few PoC compared to other neighborhoods.
We know that some people are going to use these apps in a prejudicial way, and we know some people will just go into it wanting to find a nice place to live. As long as the apps themselves aren't sold with overt or dog-whistley appeals to racism, I really don't have a problem with them.
posted by tonycpsu at 9:01 AM on February 13, 2015 [6 favorites]
More than either of those, I think it's just that the prejudice is baked into in society. Even assuming there are problems with how these factors are calculated or how they're presented, there's nothing wrong in principle with people desiring neighborhoods with low crime rates and more highly-educated residents. Still, because those variables correlate somewhat strongly with ethnicity (due to racial minorities being targeted by law enforcement and denied opportunities for many generations), there is functionally no difference between desiring neighborhoods with these characteristics and desiring neighborhoods with very few PoC compared to other neighborhoods.
We know that some people are going to use these apps in a prejudicial way, and we know some people will just go into it wanting to find a nice place to live. As long as the apps themselves aren't sold with overt or dog-whistley appeals to racism, I really don't have a problem with them.
posted by tonycpsu at 9:01 AM on February 13, 2015 [6 favorites]
You do realize these are the biggest cities in the world right? I feel you're being critical for the sake of critical, considering these cities are home to millions of people all along the social and economic spectrum.
posted by kurosawa's pal at 3:26 on February 14
Five first-world cities, all huge, is a selection issue in itself.
posted by Autumn Leaf at 9:03 AM on February 13, 2015
posted by kurosawa's pal at 3:26 on February 14
Five first-world cities, all huge, is a selection issue in itself.
posted by Autumn Leaf at 9:03 AM on February 13, 2015
I think we could at least agree there are some places safer than others; places where chances are low that you will be robbed or mugged compared to other places. While these apps may perpetuate all sorts of bad thinking, is the basic concept wrong? Wouldn't most of us like to minimize the chances of very-bad-things happening to us?
posted by cccorlew at 9:06 AM on February 13, 2015 [1 favorite]
posted by cccorlew at 9:06 AM on February 13, 2015 [1 favorite]
You do realize these are the biggest cities in the world right?
Substitute "most profitable" for "biggest" and you got it right. But that's how you get VC funding.
posted by oneswellfoop at 9:09 AM on February 13, 2015 [2 favorites]
Substitute "most profitable" for "biggest" and you got it right. But that's how you get VC funding.
posted by oneswellfoop at 9:09 AM on February 13, 2015 [2 favorites]
I know this is already a derail, but the metro areas list is fascinating. Hamburg, Stuttgart and Munich are all the same size as Berlin? As someone who was obsessed with tables in the almanac when I was a kid, this is literally mind blowing.
I'm more chagrined in a embarrassed in US-centric way that I didn't realize how large Lima and Bogota are. Lima has one-third of Peru's total population living there. Is there a comparable city in the world with that measure?
posted by 99_ at 9:37 AM on February 13, 2015 [1 favorite]
I'm more chagrined in a embarrassed in US-centric way that I didn't realize how large Lima and Bogota are. Lima has one-third of Peru's total population living there. Is there a comparable city in the world with that measure?
posted by 99_ at 9:37 AM on February 13, 2015 [1 favorite]
Berlin 3,460,725 Hamburg 1,786,448 Munich 1,353,186 Cologne 1,007,119 Stuttgart 606,588posted by benito.strauss at 9:58 AM on February 13, 2015
Autumn Leaf: Five first-world cities, all huge, is a selection issue in itself.I'm betting this website isn't depending on the number of Somalian or Ecuadorian users to make its market impact a success.
It would be much tougher to gather data on "the 5 biggest cities in the world" than "the 5 biggest cities in Western countries". It's a starting point.
posted by IAmBroom at 9:59 AM on February 13, 2015
99_: Is there a comparable city in the world with that measure?Helsinki's metro area is home to roughly 20% of all Finns. Reykjavic holds 63% of all Icelanders.
posted by IAmBroom at 10:03 AM on February 13, 2015
... which gives you an idea, comparatively, of what the land outside of Lima must be like for survivability.
posted by IAmBroom at 10:07 AM on February 13, 2015
posted by IAmBroom at 10:07 AM on February 13, 2015
I saw a marketplace.org story, there wasn't much to it, but the idea expressed was - what if machine learning big data becomes racist?
It's such troll bait that not sure what can be said about it, but I'd never considered the question before seeing the story.
posted by thoughtslut at 10:11 AM on February 13, 2015
It's such troll bait that not sure what can be said about it, but I'd never considered the question before seeing the story.
posted by thoughtslut at 10:11 AM on February 13, 2015
Is there a comparable city in the world with that measure?
Metro Seoul seems to have about half of South Korea's population.
posted by ROU_Xenophobe at 10:26 AM on February 13, 2015 [1 favorite]
Metro Seoul seems to have about half of South Korea's population.
posted by ROU_Xenophobe at 10:26 AM on February 13, 2015 [1 favorite]
benito.strauss -- the link above went by GMSA, or some similar measure -- I know there are plenty of problems with however you want to draw boundaries, but after 30 years of arguing it, I think people are able to accept in loose terms that historical boundaries are no more precise.
RE: Korea/Iceland, I assumed there would be a couple where compact geography comes into play (all of South Korea is smaller than the SD/LA/SFV corridor) -- London is at 25% as well.
posted by 99_ at 10:36 AM on February 13, 2015
RE: Korea/Iceland, I assumed there would be a couple where compact geography comes into play (all of South Korea is smaller than the SD/LA/SFV corridor) -- London is at 25% as well.
posted by 99_ at 10:36 AM on February 13, 2015
what if machine learning big data becomes racist?
If the questions you ask have implicitly racist or even just race-related assumptions, the answers you're going to get will as well.
Asking "what's the safest part of the city" is not really answerable as such, without clarifying some assumptions: chief among them, "safe for whom?" The question that they seem to be asking, though, is "what's the safest part of the city [for a typical white person whose major concern is random economic crime and not the police]" which—while not an unreasonable or irrelevant question to ask—isn't the same question. It's an easy question to answer though, with the available data.
I'd argue that where the real racism comes in isn't the asking of the question per se, but by taking the answer to that second highly-specific question and positing it as the answer to the first general one, without qualifying the assumptions.
posted by Kadin2048 at 10:46 AM on February 13, 2015 [10 favorites]
If the questions you ask have implicitly racist or even just race-related assumptions, the answers you're going to get will as well.
Asking "what's the safest part of the city" is not really answerable as such, without clarifying some assumptions: chief among them, "safe for whom?" The question that they seem to be asking, though, is "what's the safest part of the city [for a typical white person whose major concern is random economic crime and not the police]" which—while not an unreasonable or irrelevant question to ask—isn't the same question. It's an easy question to answer though, with the available data.
I'd argue that where the real racism comes in isn't the asking of the question per se, but by taking the answer to that second highly-specific question and positing it as the answer to the first general one, without qualifying the assumptions.
posted by Kadin2048 at 10:46 AM on February 13, 2015 [10 favorites]
Yeah, that's the GIGA problem innit?
posted by klangklangston at 10:58 AM on February 13, 2015
posted by klangklangston at 10:58 AM on February 13, 2015
Kadin2048, I'm not really sure how big data has material influence over us. The cloud is invisible. I wasn't thinking striclty to the topic of article (which I admit, I haven't read).
posted by thoughtslut at 11:02 AM on February 13, 2015
posted by thoughtslut at 11:02 AM on February 13, 2015
It's not really the invisibility of "the cloud" that's at issue, it's the invisibility of the bureaucratic processes within the governments that might make use of the data. To some extent, it's true that the more you hand over control to algorithms to analyze very large data sets, the more you can create situations where the machines recommend very bad (e.g. discriminatory) things, but ultimately there's supposed to be someone interpreting the results who has an understanding of whether the outputs make sense. Things like human intuition and prejudice don't disappear from the process just because we added a bunch of computers to the mix.
Is there a possibility that the decision-makers trust the outputs from the computers without understanding them? Of course. Is that any worse than simply letting people wing it without the benefit of the computers? I don't think so. We can (theoretically) review the code and the data sets that went into the computer's decision, but we can't look into a politician or civil servant's brain to tell if they're prejudiced. At the very least, the computers provide an opportunity -- not a guarantee, mind you -- that we can see the assumptions and the outcomes. That seems like a net win to me.
posted by tonycpsu at 11:23 AM on February 13, 2015 [1 favorite]
Is there a possibility that the decision-makers trust the outputs from the computers without understanding them? Of course. Is that any worse than simply letting people wing it without the benefit of the computers? I don't think so. We can (theoretically) review the code and the data sets that went into the computer's decision, but we can't look into a politician or civil servant's brain to tell if they're prejudiced. At the very least, the computers provide an opportunity -- not a guarantee, mind you -- that we can see the assumptions and the outcomes. That seems like a net win to me.
posted by tonycpsu at 11:23 AM on February 13, 2015 [1 favorite]
An amazing example of this kind of thing is the Scottish Index of Multiple Deprivation (here's an example report).
posted by yoHighness at 12:45 PM on February 13, 2015 [3 favorites]
posted by yoHighness at 12:45 PM on February 13, 2015 [3 favorites]
I saw a marketplace.org story, there wasn't much to it, but the idea expressed was - what if machine learning big data becomes racist?
Data collection and algorithms are often racist.
posted by avocet at 3:50 PM on February 13, 2015
Data collection and algorithms are often racist.
posted by avocet at 3:50 PM on February 13, 2015
I am skeptical that high concentrations of poverty and low educational attainment are not associated with a slew of quality of life issues for a neighborhood. I think our society should strive to build mixed neighborhoods with diverse populations. This seems to me to be most desirable.
posted by humanfont at 8:16 PM on February 13, 2015
posted by humanfont at 8:16 PM on February 13, 2015
"Wouldn't most of us like to minimize the chances of very-bad-things happening to us?"
Well when the "apps" play on and corroborate the irrational fear of minorities in the name of fear of *cough* "very-bad-things" then yeah, these apps would then happen to also be rubbish.
posted by stratastar at 11:51 PM on February 14, 2015
Well when the "apps" play on and corroborate the irrational fear of minorities in the name of fear of *cough* "very-bad-things" then yeah, these apps would then happen to also be rubbish.
posted by stratastar at 11:51 PM on February 14, 2015
when the "apps" play on and corroborate the irrational fear of minorities in the name of fear of *cough* "very-bad-things" then yeah, these apps would then happen to also be rubbish.
Where do you see this happening?
posted by corb at 10:47 AM on February 15, 2015
Where do you see this happening?
posted by corb at 10:47 AM on February 15, 2015
Can we not do the "you can't prove people are doing X with racist intent" conversation again?
Apps being sold to appeal to customers who want to avoid "ghettos" or "sketchy areas" are, to a large extent, appealing to people who want to avoid neighborhoods where black and brown people live. We don't have to show that people are consciously aware of this to understand that the apps will be used in this way, and to ask questions about what that means.
posted by tonycpsu at 11:03 AM on February 15, 2015
Apps being sold to appeal to customers who want to avoid "ghettos" or "sketchy areas" are, to a large extent, appealing to people who want to avoid neighborhoods where black and brown people live. We don't have to show that people are consciously aware of this to understand that the apps will be used in this way, and to ask questions about what that means.
posted by tonycpsu at 11:03 AM on February 15, 2015
But that's not what I'm responding to. I'm responding to the idea that the app is "rubbish" because itsupposedly is "corroborating the irrational fear of minorities." The app "PlaceILive" sorts for "Crime", "Police Stations", "Fire Stations" and "Prisons" for its "Safety" package. None of those things are playing on or corroborating an irrational fear of minorities. If it happens to encourage people to avoid areas with high crime, which also happen to be minority high areas, that's not corroborating irrational fear. Even the "Demographics" piece doesn't sort for race - even its "Nationality" divider only sorts for citizens "born in the US" and its "Languages" sorts for "Speak only English" - so if anything, it divides along xenophobic lines rather than racist ones.
posted by corb at 9:26 AM on February 16, 2015 [1 favorite]
posted by corb at 9:26 AM on February 16, 2015 [1 favorite]
« Older Here's what one mom wants you to know | “Hello, my name is Yusor Abu-Salha.” Newer »
This thread has been archived and is closed to new comments
Required reading (and my all-time favorite book on the subject that has never been improved upon). Anybody using "numbers can be deceiving" as a disclaimer has obviously used that book as a 'how to' manual.
posted by oneswellfoop at 7:29 AM on February 13, 2015 [1 favorite]