Become a Trustee of the Carnegie Trust
Project Title: Pilot Digitisation of Austrian Parish Books
Understanding population dynamics and inequality is a central research agenda in economics and a lot of work is being done to better understand, for example, what determines the number of children families decide to have and how these decisions are affected by big economic shocks like wars and pandemics. We have good data to study these questions in recent times. They are provided by anonymous surveys, censuses, and government records. However, if we want to understand long-run dynamics, we need these kinds of detailed data over much longer time periods.
Before statistical agencies started to assemble systematic records of a country’s population it was often churches that kept track of major life events like births, marriages, and deaths.
Recent efforts at preserving these parish books in which priests kept detailed notes of the population have led to the creation of digital archives of page-by-page photographs. This is a trove of information easily available online to the general public. To be able to use this new data source for economic analysis new ways need to be found to transcribe the data into a format that can be processed with statistical software. One recent breakthrough in this area is Transkribus, an online tool for handwritten text recognition by means of a machine learning algorithm funded by the European Commission. This pilot project is attempting to use Transkribus to systematically extract information from the photographs of parish books by digitising the priests’ handwriting and transforming it into a table structure that can then be analysed statistically.
Awarded: Research Incentive Grant
University: University of Edinburgh