Big data as a way to promote Syrian refugees' integration

We generate an astounding amount of data in our day-to-day routine, and these can prove vital in understanding various social and economic questions. With this in mind, a team at the UOC Internet Interdisciplinary Institute (IN3) decided to use big data techniques to harness mobile phone communication data from Syrian refugees in Turkey and use them to establish behavioural patterns and compare them to those of the local population. The ultimate goal is to use the information obtained to design social integration policies for refugees. The study's results, published in the scientific journal EPJ Data Science, show that the two communities behave very differently in terms of calling patterns. Given the results, the researchers feel strongly that the methodology used could be a new tool for analysing and measuring social segregation and aiding decision-making processes to fight it.

The study was carried out as part of the Data 4 Refugees (D4R) Challenge, organized by Turk Telekom and Boğaziçi University in Istanbul. IN3 researchers Daniel Rhoads, Ivan Serrano, Javier Borge-Holthoefer and Albert Solé-Ribalta, from the Complex Systems (CoSIN3) and Urban Transformation and Global Change Laboratory (TURBA Lab) groups, were the masterminds behind the study. The initiative allowed them to use call detail records from Turkey's largest mobile network operator as a way of detecting innovative solutions to the myriad of problems facing refugees living in the country.

As Rhoads explained: "Historically, social segregation between groups has been measured by taking the population's individual characteristics, such as place of residence or type of employment, into account. We wanted to explore more dynamic dimensions related to social interaction, as in the case of behaviour-based segregation. That is to say, when we look at activity patterns, is there a clear distinction between the two groups?"

For their study, the researchers chose to look at the communication patterns of refugees and locals by analysing the records of the origin and destination details of both groups' outgoing calls and SMS texts. The results show two very distinct patterns which cannot be explained by the differences in population density between the two. "We were surprised to find that local-refugee communication was practically non-existent. We figured it would be low given that the latter is a minority, but it was far lower than expected," said Rhoads.

Promoting interaction to build social cohesion

The researchers used their results to design an algorithmic solution that not only describes the current situation but also analyses possible measures to combat social segregation. The methodology is based on treating calling patterns as an indicator of refugees' behavioural traits, which can then be used to encourage refugees to relocate to areas where the behavioural patterns of local residents more closely match their own. "This approach adopts the sociological concept of homophily, which is basically individuals' tendency to associate and bond with similar others, promoting integration, not to mention social cohesion, among groups," stressed Rhoads.

Therefore, once they had identified the districts in Istanbul province where communication patterns differed between the two groups, they went on to analyse how theoretically viable it would be to set up a programme of subsidies to encourage refugees to move to places where they would have contact with a local population whose habits were more in-tune with their own. By the calculations of the study's algorithm, a monthly subsidy of €50 per family would help alleviate any possible rent increases that accompanied moving to a more expensive neighbourhood where they could integrate much better. Despite the results, the researchers are hesitant and stress that their intention is to simply quantify and study what can be expected from a policy such as this one. As Rhoads said: "Implementing a measure of this nature would be complicated; a whole range of additional factors, such as the cultural differences between communities, would also need to be considered. Integration is not one-dimensional; any number of factors come into play, both positive and negative."

Viability aside, the research has shed new light on social segregation and integration analysis by looking beyond traditional methods that study more static dimensions such as residence patterns. In Rhoads' words: "Segregation is pervasive in society, and we all recognize this, but at the same time it is very hard to define. Obviously, if two groups live in the same neighbourhood but don't interact, that's not integration. And if they do interact, but only in certain situations, that can hardly be called integration either. It's a complex issue any way you look at it, so adding another tool for measuring segregation to the toolbox can help us more fully understand this phenomenon."

The challenge of privacy

Using call detail records to obtain information on social behaviours is a recognized alternative to the traditional practice of carrying out surveys. Rhoads explained that "official national surveys are very costly in terms of both time and money. Telephone data is constantly and automatically saved, so all we need to do is interpret it."

However, he went on to say that as a source of information it also poses certain challenges, especially in a context as sensitive as the one in which refugees find themselves: "It involves designing methods that maintain study participants' privacy while also providing data that is descriptive enough to draw conclusions like the ones in our study. That has been quite a challenge."

Related articles

Rhoads, D., Serrano, I., Borge-Holthoefer, J. et al. (2020) Measuring and mitigating behavioural segregation using Call Detail Records. EPJ Data Sci. 9, 5. https://doi.org/10.1140/epjds/s13688-020-00222-1

Rhoads D., Borge-Holthoefer J., Solé-Ribalta A. (2019) Measuring and Mitigating Behavioural Segregation as an Optimisation Problem: The Case of Syrian Refugees in Turkey. In: A. Salah, A. Pentland, B. Lepri, E. Letouzé (Eds.) Guide to Mobile Data Analytics in Refugee Scenarios. Springer, Cham. https://doi.org/10.1007/978-3-030-12554-7_15

UOC R&I 

The UOC’s research and innovation (R&I) contribute to solving the challenges facing the global societies of the 21st century by studying ICTs’ interactions with human activity, with a specific focus on e-learning and e-health. Over 400 researchers and 50 research groups work among the University's seven faculties and three research centres: the Internet Interdisciplinary Institute (IN3), the eLearn Center (eLC) and the eHealth Center (eHC).

The United Nations 2030 Agenda's Sustainable Development Goals and open knowledge provide strategic pillars on which the UOC's teaching, research and innovation are built. More information: research.uoc.edu.