Mapping migrant flows with satellites
ESA tender
Bob van den Berg works at the Centre for Big Data Statistics (CBDS), which is part of CBS. ‘One of our missions at the CBDS is to harness new data sources such as big data to create new statistics. This involves working with various national and international partners to exchange knowledge and expertise.’ At the end of 2016, ESA issued a tender for a feasibility study on utilisation of new data sources, including satellite data, to uncover migration flows more quickly. ‘IT company CGI drew our attention to this tender, and we decided to pursue this project together.’ Van den Berg – on behalf of CBS – and CGI then held talks with various parties including the Dutch Ministry of Justice and Security to see which information was needed and whether these parties would participate in the project. As of that moment, Arno Sprangers – CBS statistical expert at the CBS Legal protection and security team – became involved with the project as well.
Feasibility study
Van den Berg: ‘At the start of 2017, we submitted our proposal to ESA. The project was awarded to us and we started in April.’ The project leader describes the objective of the project as twofold: ‘First of all, we wanted to see what was technically feasible based on available resources. We also held workshops together with the various parties – the policy division at the Ministry, the Immigration and Naturalisation Service (IND), police and military police – to try and ascertain what information they needed. In addition, through ESA we spoke to a number of international organisations: the International Organisation for Migration (IOM), the European Border and Coast Guard Agency (Frontex), the European Union Satellite Centre (SatCen) and the European Asylum Support Office (EASO).’
Three case studies
CBS and CGI studied the relative growth of refugee camps and migration crossroads, focusing especially on regions that draw less attention because none of the well-known (welfare) organisations are represented locally. ‘We worked out this project by way of three case studies. One historical case study in which we tried to find answers to the question: could we have anticipated the asylum flows as a result of the Syria crisis at an earlier stage? The second case study was related to recent developments in Africa. Using satellite images and social media, we looked at how the migration flows developed in Niger. The third case study focused on the refugee reception centre in Heumensoord near Nijmegen. This study enabled us to assess our approach and verify already available information with the help of satellite and Twitter data.’
Faster delivery of policy information
Bernd Burger works at the ‘Space’ division of CGI Netherlands. He is excited about the results of the feasibility study funded by ESA. ‘These show that pairing satellite and social media data generates useful information that can provide overviews of migrant flows. CGI mainly fulfills the role of information provider, Burger explains. ‘Satellites produce radar and optical images of the earth; we analyse and interpret those images. This allows us to compare optical images in time frames to one another, creating visible movements. By linking the images to Twitter and Google data and combining that with available statistical data and information from news media, for example, we obtain faster and more accurate indications of migration flows.’
Early detection
This is good news, says Burger. ‘The production of reliable asylum and migration forecasts is an extremely complex matter. Using valuable input from satellites, we would have more solid data to underpin such forecasts for the short term.’ He refers to the peak in the number of Syrian refugees in 2015. ‘That came as a complete surprise, but it had a major impact. Reception centres had to be rolled out rapidly, buildings had to be remodelled, new staff had to be hired and trained. When the flow dried up, the government was stuck with long-term contracts and empty shelters. Therefore, it is imperative that we detect changes in migration flow forecasts which could have consequences for the Netherlands and the rest of Europe at an earlier stage.’
Work for the public good
Now that the feasibility has been demonstrated, it is time for the second phase. Van den Berg: ‘That is to build a demonstration product; we are raising funds for this among various international parties at the moment. If this is successful, our final goal is to make our product operational.’ There is still some way to go, but both Van den Berg and Burger are already eagerly looking forward. Cooperation between the two partners is excellent. Van den Berg: ‘To CBS, this has been an exciting project because we work for the public good together with a commercial party. That is relatively new to us. In addition, privacy aspects must be considered, and the end product we envisage is different from the usual statistical products at CBS.’ Burger, too, is very positive about the partnership: ‘CBS is extremely competent in the analysis of data using social media and Google; they also possess a great deal of knowledge about migration and its impact on the Netherlands and on Europe.’