Link-Lives: Historical Big Data
Name of applicant
Anne Løkke
Institution
University of Copenhagen
Amount
DKK 22,538,713
Year
2019
Type of grant
Semper Ardens: Advance
What?
The aim of Link-Lives is to expand the range of register research from decades to centuries. By joining historical research methods with deep learning techniques, we will create reconstructed life-courses and family relations of almost everyone who lived in Denmark from 1787 until the modern Danish Civil Registration System was introduced in 1968. At present, this information is fragmented and unconnected. Link-Lives will be a new dynamic data system based at The Danish National Archives (Rigsarkivet). It consists of links, we will establish between information about the same person in different digitized historical sources. This will open new avenues for research with life-course and multigenerational approaches of value for historians as well as for medical and social scientists.
Why?
The biological and social life of humans is influenced by multigenerational mechanisms. However, these mechanisms are poorly understood not least due to the enormous workload which is required to establish life-courses and family relations for individuals long dead. Thus, for the era before the civil registration systems were introduced, most of our knowledge is either based on small samples of individual-level data or on nationwide aggregated statistics of variables fixed in the past. This is the challenge Link-Life can overcome by creating Historical Big Data. Link-Lives will provide easy access for researchers delivered as customized datasets from the Danish National Archives. All citizens will get access to the older part of the reconstructed life-courses through the Link-Life webpage.
How?
Danish sources to the life and health of every single individual are more complete, more diverse and encompass a longer period of time than any other country can boast. Most of this rich cultural heritage is still unexplored. We will build the backbone of Link-Lives by linking data about the same person from different censuses and burial registers. We combine manual linkage by expert historians and automatic rule based linkage methods with development of new deep learning methods. This is possible because of transcriptions contributed by numerous volunteer citizens in cooperation with Rigsarkivet and Københavns Stadsarkiv. The data structure will allow the life-courses to be enriched from multiple sources, such as parish registers, conscription records, patient records and autobiographies.