Population Trajectories Survey Methodology: Improving Coverage Error by Clustering of Freguesias and Dwelling Segmentation using Census Data in Portugal

Official statistics on living conditions and access to goods are crucial in monitoring inequality based on racial and ethnic origin and population decline as a critical contemporary demography challenge. However, data collection for minorities and sensitive information can be restricted. In this study, an independent sampling survey was designed, and machine learning algorithms were used to overcome these issues. We used clustering methods and Census 2021 data to identify essential variables and homogeneous freguesias to distribute the sample size. In addition, dwellings were segmented, and the clusters were analyzed and discussed to minimize the non-response and coverage errors. The proposed methodology provides a comprehensive final survey with proper target population overage.