Researchers use big data to derive migration indicators
A team of researchers from 10 European countries is working on ways to better understand the changing nature of migration flows and their drivers in the EU H2020 HumMingBird project. A better understanding of migration is important to develop good policies to deal with opportunities and challenges for both the host countries and the migrants. Prof. Albert Ali Salah (Utrecht University) is leading the work package on using mobile phone data to track migration patterns. “Mobile phone data can provide great coverage and address important data gaps in migration research, but privacy-aware and ethical processing of such data requires new approaches,” says Salah.
Mobile call detail records (CDR) data are stored by telecommunications companies, and carefully anonymized and aggregated indicators drawn from such datasets provide many insights related to mobility. The researchers are very aware of privacy and security risks on different levels. “We pay attention not just to the protection of individual data, but also to that of group data,” Salah explains. “For example, if you determine the mobility patterns of refugees, this may have consequences for refugees through policy changes informed by your results or influenced by media responses. It is not enough that data processing follows legal requirements, but it needs to be ethically vetted. These are quite different things.”
The collaborations between private data holders, such as telecommunication companies, and public bodies are called “data collaboratives”. Using these datasets for humanitarian purposes and for public good requires legal and ethical frameworks, as well as multi-faceted, interdisciplinary collaborations between data scientists, policy makers, and representatives of private companies.
Salah: “Through our collaboration with the biggest telecom operator in Turkey, we investigated the movement of people from Turkey towards the Greek border in March 2020, when the border was declared open. Aggregated mobile data shows the major movements, from which cities people came to the border, and what percentage of them returned to their cities afterwards.” Getting better information and deeper insights about irregular migration patterns is one of the benefits of big data processing.
The HumMingBird consortium recently released its first policy brief on addressing the international migration data gaps. In addition to mobile CDR data, the consortium explores satellite images and social media analyses, in conjunction with more traditional, qualitative research techniques, to understand the data gaps in migration. Especially considering the increased reliance on data collection by security and border agencies, understanding the data gaps and developing fair analysis approaches becomes very important for addressing gender, race, and ethnic biases in these processes.