Leveraging innovative data sources for analysis of migration patterns

The globalized world of the 21st century hosts various types of migration movements as a common phenomenon. Understanding these movements and the conditions that cause these movements is crucial as the effects of migration range from economic outcomes to the social integration of different communities. Because of this, it is common to approach this phenomenon with data-driven studies. Many studies utilize statistical and administrative data sources to study migration patterns and their demographic and socio-economic drivers. However, factors such as varying definitions of migration in available data sources and gaps between data collection periods limit the success of data-driven studies. To address these, recent studies utilize innovative big data sources. In this thesis, we propose two different studies with innovative data sources for various definitions of migration. The first study adopts a transactional data source of credit card expenditures of customers of a private bank. These geo-located transactions are employed to infer possible internal migration movements in Turkey. The latter study utilizes data obtained from Facebook to contribute to a better understanding of global migration patterns. Obtained dataset from Facebook is combined with migration stock datasets from international organizations in a visual exploratory tool. This way, the visual tool creates a medium where the innovative data source utilized can be validated. The tool is also used for visualizing the results of the case study with the transactional data source. The advantages and shortcomings of utilized innovative data sources are thoroughly discussed