Quantified Understanding of Syrian Refugee Integration in Turkey

Abstract

Turkey hosts over 3.5 million Syrian refugees. How they integrate into local communities significantly impacts the stability of the host country. In this project, we use mobile users’ Call-Detail Records (CDR) and Point-Of-Interest (POI) data to infer users’ mobility and activity patterns in order to investigate the level of integration. Using these data, we compare the spatial patterns of refugees against those of citizens. We observe a few patterns that set refugees apart, e.g., smaller travel distances, fewer high-expense activities, and separate home locations from the locals. We also establish a metric based on a citizen-refugee classifier to quantify the degree of integration. We are able to rank 11 densely populated cities, and notice that the level of integration varies from city to city. For example, Gaziantep serves as an example of a well-integrated city, whereas Sanliurfa appears to be poorly integrated.