Affiliations: [a] FEP, University of Porto, Portugal | [b] FEP, University of Porto and LIAAD INESC TEC, Portugal | [c] FEP, University of Porto and LIAAD INESC TEC and Statistics Portugal, Portugal
Correspondence:
[*]
Corresponding author: Catarina Campos Silva, FEP, University of Porto, Portugal. E-mail: [email protected].
Abstract: Luxembourg, known for its immigration history, attracts immigrants to work. This study analyses different immigrant groups in the labour market from 2014 to 2022 by using Labor Force Survey (LFS) data, Symbolic Data Analysis (SDA), and the Monitoring the Evolution of Clusters (MEC) framework. Based on the birthplace and length of residence in Luxembourg, in each year, microdata were aggregated into 21 symbolic objects. They were primarily described by 16 modal variables which are multi-valued variables with a frequency attached to each category. Moreover, clustering using complete linkage and the Chernoff’s distance was applied. The Heuristic Identification of Noisy Variables (HINoV) suggested that with just six variables, objects may be grouped homogeneously. The MEC framework traced temporal relations and transitions between the clusters, revealing some movements across the different years. Results indicate that people from the European Union (EU) and Neighbouring countries have similar profiles while the Portuguese have opposite characteristics. The Luxembourgers are somewhere in between. Profiling people from non-EU countries was challenging. The data and methodology used make it easy to replicate the work in other nations, enabling comparison of results and monitoring to continue in the future.
Keywords: Clustering, immigration, Labour Force Survey, Luxembourg, Symbolic Data Analysis
