The Research methodology of the project rests upon population data extraction, settlement identification and georeferencing of this data. The lack of any ready or aggregated Ottoman population database naturally forced me to envisage a strategy based on the collection of Ottoman primary sources that contain non-aggregated population data. Acquiring all possible sources that the project can use from the Ottoman archives in Istanbul as soon as possible was among the top priorities of the action. Archival source collection went hand in hand with data extraction and data geofencing. In accord with the predefined research strategy of the project the acquired archival documents were subjected to closer inspection and population data (male population, divided by religion) was extracted for every settlement, registered in the documents. The lack of a historical gazetteer made the process of settlement identification very difficult and time consuming, because local toponymy in Bulgaria had seen a drastic change after the national state was established. These complications were largely overcome by the usage of contemporary detailed military maps like the Russian 3-verst map (1:120 000) from 1878 and the Austrian Generalkarte von Mitteleuropa (1:200 000) from 1910 that contain in most cases the local toponymy, used in the Ottoman population registers. Accordingly, settlements’ spatial identification and the extraction of total population and house¬hold numbers from the 19th-century Ottoman population registers allowed the building of a spatially referenced database, which by 31 October 2019 contained 2254 settlements that belonged administratively to 34 sub-districts (nahiye) and 18 districts (kaza).
Each datapoint, reflecting a settlement in the 1840s population registers and the extracted Christian and Muslim population data was spatially joined with village land polygons, acquired from the Bulgarian National Agency of Cadaster. Thus, population point data from the Ottoman registers was spatially referenced to adequate territorial polygons and was analyzed further, as for instance population totals and population densities were determined on a single settlement level. The spatially joined data was aggregated further and I was able to draw realistic administrative borders of the 1840s Ottoman administrative units. The success of this task is of crucial importance for the project strategy, because the thus created administrative polygons are to be used as the primary analytic spatial unit of POPGEO_BG in comparing population fluctuations over large time periods. The approach in which Ottoman point data was spatially joined to village land polygons, the polygons belonging to the same administrative sub-district (Ott. nahiye), were further aggregated to draw the borders of the sub-districts and finally the so-formed sub-districts were once more aggregated to draw the borders of the districts (Ott. kaza), proved to be very successful. I was able to generate 29 territorial units, which occupy an area of 38 537 square km, which constitutes roughly about 35% of the territory of Bulgaria. Creating POPGEO_BG primary units for spatial analysis allowed me to run multiple tests with spatially referenced population data that I have built, thus creating larger segments of the country’s population densities in 1840s.
