Using Internet data to measure migrations

The Information Science Colloquium series welcomes computational sociologist Bogdan State of Stanford University, where he is Honors Cooperative Program student in the master's of computer science program. He worked previously as a data scientist for Facebook. For more information on his research, visit his LinkedIn page.

Title: Using Internet data to measure migrations

Abstract: Measuring human migrations remains a very difficult problem for social scientists. Data collected through traditional means is oftentimes outdated, inconsistent, or altogether absent. The data revolution offers an excellent alternative for the measurement of migrations. In the future policy makers and social scientists could be relying on new sources -- IP geolocation, call-record and GPS data -- to detect human migration patterns in real time. These new data sources come with their own problems, however, and new methods are needed to remove both noise and bias. I will present work my coauthors and I did towards this goal, focusing on the measurement of migrations using IP geolocation, as well as on GPS data. Solving inherent measurement issues is important not only for supplementing censuses, however. Building a thorough methodology for estimating migrations also opens new possibilities. As examples, I will present research focusing on hard-to-reach populations such as highly-skilled migrants, as well as a study that investigates intergroup relations. These examples underpin the argument that human mobility is one of the areas where the computational turn has enormous value to bring to both science and policy.