Digital trace data presents an opportunity for promptly monitoring shifts in migrant populations. This contribution aims to determine whether the number of European migrants in the United Kingdom (UK) declined between March 2019 and March 2020, using weekly estimates derived from the Facebook Advertising Platform. The collected data is disaggregated according to age, level of education, and country of origin. To examine the fluctuation in the number of migrants, a simple Bayesian trend model is employed, incorporating indicator variables for age, education, and country. The Facebook data indicates a downward trend in the number of European migrants residing in the UK. This result is further confirmed by the data from the Labour Force Survey. Notably, the outcomes reveal that in the run-up to Brexit, the most significant decline occurred among the age group of 20 to 29 years old – the largest migrant group – and the tertiary educated. This analyses could not be implemented with traditional data sources such as the Labour Force Survey, because this level of disaggregation is not provided. However, there are also important limitations associated with digital trace data, such as algorithm changes and representativeness. These limitations need to be addressed by employing sound statistical methodologies. Nevertheless, this research shows the potential of digital trace data in anticipating migration trends at a timely granularity and informing policymakers.

中文翻译：

---

通过数字痕迹评估及时的移民趋势：英国脱欧前的案例研究


数字追踪数据为及时监测流动人口的变化提供了机会。本贡献旨在利用 Facebook 广告平台的每周估计数据来确定 2019 年 3 月至 2020 年 3 月期间英国 (UK) 的欧洲移民数量是否有所下降。收集的数据根据​​年龄、教育水平和原籍国进行分类。为了研究移民数量的波动，采用了一个简单的贝叶斯趋势模型，其中纳入了年龄、教育程度和国家等指标变量。 Facebook 数据显示，居住在英国的欧洲移民数量呈下降趋势。劳动力调查的数据进一步证实了这一结果。值得注意的是，结果显示，在英国脱欧前夕，最显着的下降发生在 20 至 29 岁的年龄组（最大的移民群体）和受过高等教育的人群中。这种分析无法使用劳动力调查等传统数据源来实施，因为没有提供这种级别的分类。然而，数字跟踪数据也存在重要的局限性，例如算法变化和代表性。这些限制需要通过采用合理的统计方法来解决。尽管如此，这项研究显示了数字追踪数据在及时预测移民趋势并为政策制定者提供信息方面的潜力。