Financial Datasets: Leveraging Transactional Big Data in Mobility and Migration Studies

Abstract

In the Globalised world of the 21st century, all types of human migration are a widely encountered phenomenon with many effects on individuals and societies in areas ranging from economic outcomes to social integration of migrants in the receiving regions. With the increasing availability of various big data sources, many studies aim to incorporate traditional statistical and administrative data sources for migration flows with innovative data sources. The introduction of innovative data sources enables researchers to address challenges from various definitions of migration with different scales. Our study utilises a transactional dataset of a private bank in Turkey to infer possible internal migration movements of individuals. We aim to distinguish customers who permanently settled in different cities by defining constraints on transaction patterns of individuals from a sample dataset with one year of transaction coverage. Even though the presence of limitations discussed in the study, our results demonstrate that such datasets can be employed for analysing migration patterns. Additionally, utilising transactional data also allows us to analyse the spending behaviours of individuals in their origin and destination cities. Our findings suggest that individuals in our sample tend to narrow down their spending in terms of categorical diversity and amount diversity, while they slightly increase the number of locations they visit after relocating to a new city.