The workshop was co-organized by the EHPS-Net and the Edinburgh University in Edinburgh within the BioQuarter building of the Administrative Data Research Network Scotland (ADRC-S).
The main purpose of the workshop was to discuss, compare and develop methods and standards for automatically geocoding historical events/records and deriving research uses of historical ‘small area’ geographical variables. These methods are important for a wide range of applications within social science, geography and epidemiology, and are particular crucial to historical demography. Previous workshops were devoted to similar topics but considering methods and standards for storage, integration and visualization of data with multiple spatio-temporal representations.
Despite the importance of space in longitudinal demographic analysis, few spatial analyses have been carried out. The main reason is the lack of geographical information linked to the historical demographic data. Having access to this information may substantially benefit the research. Geographical information is thus important to consider when developing the IDS. To enable the linkage between IDS databases and other geographical information sources, we need to store geometrical representations of geographic objects in IDS. This could for instance be storing a geographic representation of buildings and property units. When individuals are linked to these buildings and property units it will be possible to investigate the environment where these people lived.
Several projects across Europe are working (or plan to work) with linking digitized geographical objects from maps with geographical locations in demographic databases. One method has been to use text labels in the historic maps as well as supplementary text sources to these maps. A limitation with the current solution is that the connection between the demographic and geographic data has only been made on an aggregated level (in terms of individuals). To enhance the possibilities for analysis, a connection on individual level (that each person in the demographic databases is linked to one or several buildings or property units) is also needed. However, this linkage requires methodological development because the two sources by nature often use different time representations. The historical maps can be regarded as snapshots of the conditions at a certain time whereas the demographic database is using continuous object lifelines (a person exists from birth to death).
There were 9 presentations by different groups from Europe and United States. Several of these presentations dealt with spatial analysis of demographic and socioeconomic data: infant mortality in Madrid, historical vital events in Edinburgh and immigration in New York City. Other group of presentations where focused on methodologies for building large GIS infrastructures on a semiautomatic system, like the case of Ian Gregory and his work on methodologies for analysing unstructured texts—including large corpora of historical books, periodicals and official reports—within a GIS environment. Finally, the presentations on the Digitising Scotland, the MESH and the ALPAGE projects as well as the implementation of a Historical Gazetteer in Kiel, Germany revealed the benefits and limitation of open source tools and open data usage for the historical GIS studies.