This is fantastic! We were just noticing how segregated our own neighborhood near Boston is, and wondering what drives that and what could be done about it. You can see the same thing all over the city -- neighborhoods that are much more white than average right next to neighborhoods that are much less white than average.<p>Really interesting that this could be self-generated with very little bias (setting aside that there's definitely still some intentional housing discrimination in Boston). And really interesting that it could potentially be reversed if people started to avoid neighborhoods that are highly segregated in their "favor."<p>I wonder if integration could be advertised as a benefit of certain properties on real estate sites like Zillow. What would happen if home listings had a "well integrated neighborhood" indicator for neighborhoods that have about the same racial balance as the larger area, the same way they have indicators for good schools and public transportation and so on? Would that be appealing to actual buyers the same way it's appealing to the Polygons in the model?<p>The risk is that an index like that could be used to <i>encourage</i> segregation instead -- but I'm hopeful that, on average, we're better than that at this point.<p>Here's one census map if you want to check out your neighborhood:<p><a href="http://www.socialexplorer.com/" rel="nofollow">http://www.socialexplorer.com/</a><p>You can show racial data under "Change Data." We also found it helpful to change "Show data by: Tract" to "Block group" (more fine-grained), and to use quantile cutpoints under the color palette menu.
I was surprised to find that this was partly done by the incomparable Vi Hart. But why should that have been a surprise: she has a distinctive knack for presenting mathematical concepts in a way that makes them understandable.
This one is incredibly fascinating, even though it is simplified. One thing to keep in mind is that even if you exclude racial biases, you will maintain segregation if people have other - correlated - biases or limitations on their ability to move:<p>In many countries, poverty is highly correlated with race, for example. This is certainly the case with the US, but also elsewhere. I live in London, and you see interesting effects of this.<p>The inner city boroughs are quite segregated, both by race and wealth.<p>Meanwhile, some of the outer boroughs are showing the reverse effect, where property price points appears to be a driver for mixing. E.g. Croydon, where I live, is one of the most mixed in London - it's at a price point that creates both young professionals of all races, and more established families of all races who are united in finding the inner boroughs either too expensive or too poor.<p>But overall: Imagine that nobody had a racial preference, but had a wealth preference - and limitations.<p>Now to overcome segregation, you face a near insurmountable barrier: Wealthier people would need to be willing to settle for housing and an environment of a much worse standard than they can afford, and poorer people would be unable to find housing that makes much difference.<p>This is one of the biggest problems. The recent US situation with demonstrations over police killings, the race aspect has been blown out of proportion: You don't solve anything by focusing on the race issue, because so much of the violence is correlated more strongly with poverty than with race. You want to solve racial issues, start by addressing poverty. You'll <i>still</i> have race issues at the end of it, but it will turn out vastly smaller than what it appears, and you'll have removed a substantial source of excuses for racial biases.
Wow! Not only an educational game about social diversity, but also: agent-based models, phase transitions, irreversible thermodynamics. And, last not least - cute shapes.<p>I am definitely adding it to my educational game recs: <a href="https://hackpad.com/Science-based-games-J0X4MSberlM" rel="nofollow">https://hackpad.com/Science-based-games-J0X4MSberlM</a>
I wanted to point out one of their conclusions in particular:<p>> <i>1. Small individual bias → Large collective bias.</i><p>> <i>When someone says a culture is shapist, they're not saying the individuals in it are shapist. They're not attacking you personally.</i><p>This is a really important point that gets a lot of arguments hung up on nonproductive "not all men"-type fooforaw lately. If someone says that e.g. gamer culture is sexist, or even if they (perhaps inadvisably) use shorthand and say "gamers are sexist," they are not saying that I and every other single person who plays games is personally sexist.
Beautiful!<p>For the math and game theory links, make sure to go all the way to the bottom. The post is based off the work of Nobel Prize winner Thomas Schelling and the style is inspired by Bret Victor's Explorable Explanations. There's more in the post, of course.
On the third board (the first one with a slider), setting the bias to 100%, segregation stays under 10% (I've run it for a bit and the max was 8%). Obviously we need to tell everyone be act like "I'll move if less than 100% of my neighbours are like me"! ;)
I enjoyed playing with the graphs and everything, but I question whether this model has much relevance to the real world. Is there a strong reason to believe that these effects would survive a model of "I want to move" that is not solely based on "too many people unlike me live near by" and/or "not enough people unlike me live near by"? Indeed, is there a strong reason to believe that a binary modality of "I'm happy/unhappy," (the post gestures in the direction of a third mode, "I'm neither happy nor unhappy," but in fact in their simulations that third mode is indistinguishable from "happy") is a good abstraction of people's moving decisions?<p>The data paper they posted a link to suggests that there is unlikely to be an equilibrium, contra the message of this post.<p>It seems like it's more an explanation of a mathematical model and a prescriptive political position, rather than a description of anything real in society.
Great post, though I'm missing a definition of how do they mathematically define segregation. (Checking from the JS code, it seems to be an average over the shapes of the proportion of neighbor shapes which are like them.)<p>Also, the title confused me a bit as I expected something related to Flatland.<p>By the way, one of the two people behind this is Vi Hart, which has also made some brilliant videos featuring math, music, and brilliant silliness: <a href="https://www.youtube.com/watch?v=4niz8TfY794&list=UUOGeU-1Fig3rrDjhm9Zs_wg" rel="nofollow">https://www.youtube.com/watch?v=4niz8TfY794&list=UUOGeU-1Fig...</a>
> it's about deciding what we want the world to look like, and settling for no less<p>Perhaps it's about accepting people for who and what they are, even when they fail to satisfy your <i>own</i> biases, and building an understanding of the social world on that basis -- rather than being about privileging your opinion of how the world ought to be over your perception of how the world <i>is</i>, and bridging the gap between there and your thesis via the <i>post hoc</i> fallacy.<p>...nah. That would be stupid.
Beautiful UI. Schelling's models of segregation is a great example of how simulation and data-vis can create insights in the humanities & social sciences.<p>I made a similar (though much less beautiful) simulation for a workshop in in computational philosophy last semester - but replaced intentional with random movement when uphappy.<p><a href="http://dave.kinkead.com.au/models-of-segregation/" rel="nofollow">http://dave.kinkead.com.au/models-of-segregation/</a>
Problem with this model. If you turn the bias way up (demand 80-90 percent of your neighbors be like you) the segregation drops to zero. That doesn't seem to make sense and I think it's because they're moving to random locations that have low probability of being "better" by the binary nature of the metric.
This is a fantastic visual demonstration of a simple, but surprising (if you haven't seen it before) effect! Congrats to Vi Hart & Nicky Case (I presume Nicky did the web/javascript side of things?).<p>This isn't <i>just</i> about race / skin colour either - you can see exactly the same effects in things like political viewpoints where people also tend to segregate into areas or simply social circles of people that share their own preferences. (pace the praphrased quote: "How could politician X win the election? <i>Nobody</i> I know would vote for them!")
I wonder how we could maybe put this idea to work as part of immigration reform or the much-discussed work visa topic. After all, moving is tough so not everyone can just up and move for diversity's sake, but if someone is already moving, maybe a requirement should be that they move to communities unlike them?
What does it say that anything over 75-85% "segregationism" leads to an almost completely desegregated community of shapes over time (though, they are <i>constantly</i> moving)?
There are many settings for which this process doesn't converge. I consider it a feature rather than a bug.<p>For example, if you have the parameters at 33% and 100%, you converge on a racist/segregated state. However, at 33% and 95%, people keep moving around. You see the large contiguous regions that would otherwise be suburban enclaves hollow out. But you never get to a stable state, which is kinda like real life so it works.<p><i>Small individual bias → Large collective bias.</i><p>This is so incredibly true. I wrote about this. The blog post is specific to VCs, but I think the Racist Judges Problem is much more general. (<a href="http://michaelochurch.wordpress.com/2013/10/08/vc-istan-2-the-racist-judges-problem/" rel="nofollow">http://michaelochurch.wordpress.com/2013/10/08/vc-istan-2-th...</a>)<p>Let's say that we're having a cat beauty contest and, for the sake of argument and egregious simplification, that 35 states are completely non-racist and 15 states are extremely racist against orange cats. Let's also assume that beauty is uncorrelated to kitty-race and that there are only two colors: white (80%) and orange (20%). You'd expect that 7 contestants would be orange: a proportionate share from the non-racist states and none from the racist states. But you'll actually get very few, because the non-racist states still want their cats to win, so they'll be <i>de facto</i> racist for strategic reasons except for an occasional "weird" state (say, Minnesota) that nominates an orange cat and Fox News doesn't shut up about it being "political" in its "making a statement" (even if the orange cat from Minnesota <i>was</i> the most beautiful). Thus, when you look at the contestants and see 49-50 white cats, it creates a horrible and completely false perception that white is the standard of cat beauty and that no one finds orange cats beautiful.<p>In other words, small differences in preference, once they gain a certain social currency and acceptability, snowball into something a lot more horrible. This is also why it's so important to keep racism socially unacceptable, and why Stetson Kennedy's infiltration of the Ku Klux Klan (he exposed it as childish and ludicrous) was so powerful.
It plays with many variables to show the consequences of these choices, but never questions the basic assumption that shapes are happier in a diverse environment.<p>It then goes on to conclude "Demand diversity near you".<p>I know what they try to accomplish, but I still think it's a dishonest (or at least incomplete) approach.