Geo-referenced database for the architectural heritage: a tool for the study and enhancement of the medieval churches in the municipalities of southern Albania

The municipality of Finiq, as well as the adjacent ones - Delvinë, Sarandë, Koni- spol, and Dropull - is populated by a large group of sacred buildings dating back to the early Middle Ages. We define the architectonic style of these churches as Byzantine, although actu- ally the component closely related to the architecture of Byzantium is flanked by construc- tional, spatial, and decorative types peculiar to the Epirus territory. The specificities of this built heritage are mainly due to the geographical location of Epirus as a frontier site between the Latin, Serbian, and Greek areas of influence, at a time of the diaspora of the Byzantine legacy after the fall of the capital of the Empire to western hands in 1204. Currently, this set of sacred architecture, mostly monasteries and hermitages set in scenically significant locations, are being rediscovered, both from the scholarly and the purely touristic point of view. What seems to be lacking at the moment is a tool capable of relating the buildings and through this relationship, which can be made explicit through the analysis of constructive, spatial, and stylistic features, promoting critical knowledge among scholars, also with a view to more controlled and conscious actions of protection, and foster their success in the sphere of valorization and promotion. Such a tool can be configured through a georeferenced database, capable of storing within it data related to direct sources (photographs, two-dimensional processing, models, and 3D point clouds) and indirect sources (documents, bibliographic citations in scientific literature, ancient and modern cartographies). Systematizing this mass of information permits the scholar who accesses it not only a more structured knowledge of the object of study but also an infinitely greater possibility of relating the building studied to other buildings in the area simply by analyzing the relationships that are automatically created within the database. A relevant example may be the emergence of the Monastery of St. Nicholas in Mesopotam, which presents peculiarities from the points of view of spatial and constructive typology - a single hall with a narthex, four domes supported by a central column, two apses - and decora- tive apparatuses both sculptural and pictorial. The study of these features, not yet fully under- stood historically and culturally, can be facilitated by a comparative tool such as the database under study. In the future, such a formed relational structure can be the basis for further insights that implement artificial intelligence and machine learning processes in order to automate some steps of the cognitive process. Indeed, we know that, given a quantitatively relevant input data, a classification algorithm can be trained and thus can perform tasks that, if tackled by human operators, would be excessively time-consuming. A classification algorithm could be trained to recognize recurring decorative motifs, spatial proportions that deviate from the trend, structural solutions particularly related to a historical era, etc. This tool, in addition to aiding the construction of more structured knowledge on the specific topic, could be a stan- dard of acquisition-analysis-results applicable to other areas of built heritage.