Archaeology is a source of essential data regarding the nature of human societies. Researchers across the behavioural and social sciences use archeological data in framing foundational arguments, and archaeological evidence also frequently undergirds debate on contemporary issues, such as the rise of social inequality or persistence of certain societies over others (Kintigh et al. 2014). Our ability to address fundamental questions such as the evolution of human culture, and waxing and waning of cities and civilisations, depends on our capacity to acquire, manage and analyse large evidentiary datasets. Current attempts to synthesise markers of complexity suffer from the problem of being based on interpretations (e.g., Seshat: Global History Databank asks researchers to classify societies in their area of expertise but does not require raw data for these claims), or being created manually and thus scale poorly (e.g., Social Reactors Project). While crowd-sourced projects such as Seshat (Turchin et al. 2018) have value, their derivative nature is not self-correcting and can suffer from confirmation bias and paradigmatic lock-in. Meanwhile, individual smaller-scale studies, whether archaeological or historical (Grove 2011; Ortman et al. 2015; Hanson and Ortman 2017), utilise primary data and combine formal synthesis of 'raw' data with falsifiability. Manual examination of primary data is, however, time and labour intensive.

The Social Dynamics of Ancient Mediterranean (SDAM, www.sdam.au.dk) project explores the challenges of computer-assisted aggregation and streamlining of archaeological and historical data in several quantitative case studies focused on the ancient Mediterranean. Before we commence cross-regional comparison, we need to tackle problems inherent in archaeological data. These include temporal, spatial and attribute uncertainty or inconsistency, diverse cross-regional standards and incompatible datasets. Our aim is to develop generalised analytical workflows and approaches that allow us to work directly with raw archaeological data. Once we can compare and contrast trends in different categories of data with less effort and time and in transparent and reproducible manner, then we can expect to shed more light on macro-historical questions concerning the dynamics of human societies in general.

In our joint presentation, we will present our digital collaborative environment and demonstrate how we deal with some of the issues. In particular, we will focus on the use of temporal and spatial aspects of large-scale epigraphic datasets as a proxy for tracking changes in the Roman transportation network over time. To analyse these changes, we will subsequently demonstrate the applicability and usability of algorithms for formal analysis of dynamical networks.