The entry of immigrants into their incoming country as well as their subsequent spread over the country's territory occurs through processes that have a strong spatial bias. The specific migration groups (here defined as the set of immigrants with the same country of origin) tend to reveal different levels and patterns of their spatial concentrations.
The understanding of these similarities and dissimilarities in spatial behaviour is important task from both academic and policy perspective. This study has a quantitative character and it is based on the study of "spatial relatedness" of migration groups, which is defined as a rate, how much do given migration groups concentrate into the same regions.
The study is based on a simple assumption that the spatial relatedness mirrors mutual proximity or similarity between these groups in other respects (e.g. cultural or economical). Plenty of datasets describing spatial distribution of migrants within 34 territorial systems on various scales (from global to local) are analysed.
These systems will be analysed separately at first and then a synthesis of these partial results will be done with the aim to discover prevailing pattern of the spatial relatedness of migration groups. The assessments consists of several steps, which include particularly calculation of the spatial relatedness of migration groups, identification of patterns of these relatedness using network and cluster analysis, and testing their basic causations by regression analysis.
Then, the thesis presents and tests a method utilizing the spatial relatedness approach for improving forecasts of spatial distribution of individual migration groups. Finally, the ways how the methods introduced in this study can be utilised are presented on the example of the Czech territorial system.