Seminar Developing scalable and standardized ai tools for register-based administrative data collection, organization, and analysis for the implementation of large longitudinal population registers

Speaker: Diego Ramiro-Fariñas (Institute of Economics, Geography and Demography, CSIC, Spain)

Abstract:

Historical administrative and statistical sources—such as censuses, civil registers, and mortality records—represent foundational materials for demographic research. While the value of these sources is well established, their large-scale reuse remains constrained by issues of format, legibility, and structure. Depending on the historical period, sources range from fully handwritten and unstructured formats to semi-structured forms with pre-printed templates manually completed by census agents. Even in the latter case, variability in handwriting, ink types, and layout inconsistencies introduce significant obstacles for automated processing. This presentation addresses these limitations through the development of scalable, AI-based tools for the transcription and structuring of register-based administrative data, aiming to support both historical reconstruction and statistical reuse for the implementation of Large Longitudinal Population Registers.

Organizer: Francesco Scalone