Measuring and forecasting migration with traditional and novel data

Description

Having up-to-date information about the nature and extent of international migration is important for policymaking, such as labour market policy or social services, as well as for prediction and planning. However, timely and reliable statistics on international migration are often imperfect, incomparable or missing. With the rise of the Internet, new data sources offer opportunities to complement traditional sources for migration statistics. In particular, the availability of large quantities of data derived from social media and from search engines has opened new avenues of research.

In this talk, three examples of combining traditional and novel data to measure and forecast migration are presented. The first one demonstrates how official data on migrant stocks derived from censuses, population registers, Labour Force Survey and Facebook Advertising Platform can be integrated to produce harmonised estimates and forecasts of bilateral migrant stocks within the European Union. The second case study is on the use of data from Twitter to capture the effects of the COVID-19 pandemic on internal mobility within India in the absence of migration data. The third example shows the use of Google Trends data to make short-term forecasts of migration from Romania to the United Kingdom.

A common denominator of the three examples is the use of Bayesian inferential framework to integrate data from various sources. Bayesian inference produces estimates and forecasts of quantities of interest, such as the harmonised count of migrants, by combining data described by a statistical model with prior knowledge about the model parameters expressed through prior probability distributions.

Period	28 Nov 2022
Held at	Sun Yat-sen University, China
Degree of Recognition	Local