評論
打印
Among many correlations detected, the researchers found that high levels of B vitamins
在檢測到的很多相關性中，研究人員發現高濃度的B族維生素
(abundant in red meat, whole grains and dark leafy greens),
（在紅肉、全谷物和深色綠葉蔬菜中含量豐富）、
alcohol and caffeine were associated with high-rent districts, generally in the middle of cities,
酒精和咖啡因與高租金的地區相關，這些地區通常都位于城市中心，
while opioids and antidepressants went hand in hand with an abundance of people with low-skilled occupations.
而阿片類藥物和抗抑郁藥與大量低技能工作者相關。
These general correlations are not, perhaps, that surprising (though another,
或許這些普遍的相關性并不令人驚訝（雖然另一種相關性，
between a lack of internet connection and an antihypertensive medication called atenolol, is intriguing—
缺乏網絡連接與一種叫做阿替洛爾的降壓藥之間的聯系，很有趣——
possibly being mediated through people's age).
可能會隨著年紀調節）。
But Dr Samanipour found that in combination they could predict quite accurately the demographic make-up of a sewage plant's catchment area.
但Samanipour博士發現，結合起來看，它們可以相當準確地預測污水廠集水區的人口構成。
He did this by taking the model he and his colleagues had devised, based on their 100 initial plants, and applying it to nine others.
通過利用他和同事設計的模型，以他們最初的100個處理廠為基礎，并將該模型應用到其他9個處理廠中。
For each of these they forecasted the socioeconomics of the surrounding area and then tested those forecasts against reality, as revealed by the census.
他們對每一個周圍區域的社會經濟進行了預測，并將這些預測與人口普查披露的實際情況進行對比。
These forecasts were quite specific: how many people in an area had finished secondary school;
這些預測相當詳細：這片區域內有多少人中學畢業；
how many had cars; how many were single parents; and so on.
多少人有車；多少人是單身父母等等。
Of the 37 questions asked by the census about education, occupation, income, social life and housing,
在人口普查調查的37個關于教育、職業、收入、社交生活以及住房的問題中，
the model was able to predict the pattern of answers to 30 of them with 25% or less deviation from the values calculated by the census.
該模型能夠預測其中30個的答案模式，且與人口普查計算值的偏差不超過25%。
Based on these results the researchers argue that, considering the relative ease,
基于這些結果，研究人員認為，鑒于收集和分析
low cost and high frequency with which wastewater samples can be collected and analysed, their model could profitably be used,
這些污水樣本的容易度、低成本和高頻率，他們的模型可以被有效利用，
if not as a replacement for censuses, then at least as a way of keeping track of changes in the characteristics of populations more or less continuously,
即便不能替代人口普查，至少也可以作為一種持續追蹤人口特征變化的方法，
as they happen, rather than in a punctuated manner every five or ten years.
而無需間斷式的每五年或每十年一次。
譯文由可可原創，僅供學習交流使用，未經許可請勿轉載。
