来源：《新科学家》
原文见刊日期：2022年11月5日
本文适合2024考生
People living in large cities are more segregated, and mix less with those from different socioeconomic backgrounds, than people in small towns, according to an analysis of anonymous phone data from 9.6 million people in the US.
一项对美国960万人匿名电话数据的分析显示，与小城镇的人相比，大城市的人更加隔离，与来自不同社会经济背景的人相处更少。
A long-standing premise of urban design is that cities encourage interactions between different economic groups and so lead to less segregation. One common way to measure this is to look at where people live and their corresponding economic status.
城市设计的一个长期前提是，城市鼓励不同经济群体之间的互动，从而减少隔离。衡量这一点的常见方法是观察人们居住的地方和他们相应的经济状况。
Measurements like these seem to support the idea of cities enabling mixing, but some argue this static snapshot is incomplete and needs to include how people interact when they aren’t at home.
像这样的测量结果似乎支持了城市实现互动的观点，但一些人认为这种静态快照是不完整的，需要包括人们在不在家时的互动。
Now, Jure Leskovec at Stanford University and his colleagues have used smartphone data on 1.6 billion interactions between 9.6 million people in 382 US towns and cities to show that the largest cities in fact cause more segregation between people from different socioeconomic backgrounds.
现在，斯坦福大学的朱尔·莱斯科维奇和他的同事利用智能手机数据，对美国382个城镇960万人之间的16亿次互动进行了研究，结果表明，事实上，在最大的城市中，不同社会经济背景的人之间存在更多的隔离。
First, Leskovec and his team looked at where each phone, representing a person, was located at around 2am, when they were likely to be at home sleeping. “From this, we infer their income, their socioeconomic status, and then we look during the day, as these cell phones move around, how often these cell phones cross paths,” says Leskovec.
首先，莱斯科维奇和他的团队观察了每个代表一个人的手机在凌晨2点左右的位置，这个时候人们可能正在家里睡觉。莱斯科维奇说：从这些数据中，我们推断出他们的收入和社会经济地位，然后在白天，当这些手机移动时，我们观察这些手机路径交叉的频率。”
The researchers defined crossing paths as two people being in a 50-square-metre area within a 5-minute period. They tried other definitions, changing the space and time boundaries, but it didn’t alter the findings.
研究人员将交叉路径定义为两个人在5分钟内出现在一个50平方米的区域内。他们尝试了其他的定义，改变了空间和时间的边界值，但这并没有改变研究结果。
They found that their interaction-based measure of segregation was 38 percent lower, averaged across all areas, than a static measure of segregation, indicating that studies may have overestimated segregation when only considering home location. But they also found that segregation was 67 percent higher in the 10 largest cities they looked at than in the smallest, defined as having fewer than 100,000 people.
他们发现，在所有地区，基于互动的隔离度量比静态隔离度量平均低38%，这表明，当只考虑家庭位置时，研究可能高估了隔离。但他们也发现，在他们观察的10个最大的城市中，隔离的程度比人口少于10万的小城市高67%。
Leskovec thinks this is because cities have so many places to go. “This basically allows people to kind of bucket themselves into more fine-grained buckets. The more buckets you have, the more equals to yourself you can find because you slot yourself in the appropriate bucket,” he says.
莱斯科维奇认为这是因为大城市有太多的地方可以去。“这基本上可以让人们把自己归类到更细致的分类中。你拥有的分类越多，你就越能找到与自己相称的东西，因为你把自己归入了合适的分类。”
The least-segregated locations were places like museums and shopping centres that people frequently visit. One way of reducing divisions between people of high and low income could be to build such places between diverse neighbourhoods to act as a “bridge” between two groups, says Leskovec.
隔离程度最低的地方是人们经常光顾的博物馆和购物中心。莱斯科维奇说，减少高收入者和低收入者之间的分隔的一种方法可能是在不同的居住区之间建立这样的场所，作为两个群体之间的“桥梁”。
“What is new and interesting in what they’re doing is the use of mobile phone data,” says David Manley at the University of Bristol, UK. “That large kind of spatial data set means that they’re able to capture the population movements in a way that we haven’t been able to do before.” However, the data set isn’t detailed enough to say whether people are interacting or just being in the same space, he says, and different data sets could be needed to get that level of detail.
英国布里斯托大学的大卫·曼利说：“他们所做的事情的新奇和有趣之处在于手机数据的使用。如此庞大的空间数据集意味着，他们能够以一种我们以前无法做到的方式捕捉到人口流动。”然而，他说，数据集还不够详细，不能说明人们是在互动，还是只是在同一个空间，可能需要不同的数据集来获得这种详细程度。