Methodological Approaches for Multidimensional Personal Data Creation
Methodological Approaches for Multidimensional Personal Data Creation

Summary/Abstract: This paper provides information on the description of metadata when using an algorithm to generate a multidimensional synthetic dataset. And addresses the challenges associated with collecting and using extensive datasets for scientific research, particularly in the context of sensitive information governed by legal frameworks such as the GDPR and the Bank Secrecy Act. The methodology under consideration employs simulation techniques to create a dataset comprising 36 distinct variables categorized into demographic, personal, and banking characteristics. This synthetic dataset is essential for empirical studies where data availability is restricted due to legal constraints. The research draws on diverse data sources, including the Bulgarian Census 2021, the National Statistical Institute, and the Bulgarian National Bank, ensuring comprehensive coverage for deriving the distributions. We emphasize the importance of validating the generated data to meet quality standards and support effective modeling. This study contributes to the ongoing discourse on data synthesis in data science, highlighting innovative strategies for addressing data shortages while at the same time following Eurostat's best practices for describing metadata, by making a detailed breakdown of all variables and analyzing the need for their inclusion in the summarized set of information, in view of the objectives of the study.

• Details
• Contents