Laboratory
Migration and Mobility
At a Glance
Projects
Publications
Team
Project
Estimating Migration and Mobility after Natural Disasters
Emilio Zagheni; in Collaboration with Monica Alexander (University of Toronto, Canada), Kivan Polimis (University of Washington, Seattle, USA)
Detailed Description
This project complements social media data with official statistics and surveys in order to monitor population change after natural disasters, such as hurricanes or floods. While demographic data collected from social media contain known biases and issues of nonrepresentativeness, they have the advantage of being available almost in real time. With appropriate statistical corrections, they provide a valuable source of information to produce timely estimates and improve migration theory.
In a particular case study, we used data from Facebook’s advertising platform to estimate out-migration from Puerto Rico in the months after Hurricane Maria. Using the difference-in-difference method in combination with data from the American Community Survey, we have found that in the three months after the hurricane hit, approximately 180,000 people moved out of the island to the USA, mainly young people between 15 and 30 years old. In addition, we have seen that these young adults moved to US states with a large Puerto Rican community: In Florida, New York, and Pennsylvania, the population of Puerto Ricans increased the most. Conversely, older people were more likely to stay in Puerto Rico when considered in terms of both absolute and relative change. We have also found evidence of subsequent return migration to Puerto Rico over the period January 2018 to March 2018. These results illustrate the power of complementing social media and traditional data to monitor demographic indicators over time.
Compared to data sources such as flight passenger data, mobile phone data, or change of address and school enrollment, anonymous Facebook data that are routinely produced for advertisers offer valuable demographic information about sex and age because it is possible to assess not only how many people moved place but also their demographic characteristics.
This project builds on previous research that used social media data in demographic research. We have been collecting data over an extended period to measure migration changes in response to shocks such as a natural disasters. The richness of the data available allows us to consider changes in migration by age, sex, and destination, and the data are not limited to specific events. As we extend our models, we expect to be able to assess the differential migration and mobility responses to a number of shocks, including disease outbreaks and other types of crises. This information will inform our theoretical understanding of migration after sudden and unanticipated events, including the role of social networks of support, the spread of information, and distance measured along various dimensions.
Data and Surveys, Migration
Publications
Klein, J. D.; Weber, I. G.; Zagheni, E.:
Demography, 1–19. (2024)
Alexander, M.; Polimis, K.; Zagheni, E.:
Population and Development Review 45:3, 617–630. (2019)