The narrative of the human history concerns contrast and change or their absence. Understanding the various historical trajectories followed by human populations subjected to a multiplicity of circumstances is the ultimate goal behind the study of the human past. In this lies the opportunity of characterizing causal mechanisms shared by both past and present societies, which can provide benchmarks for future policymaking decisions. The challenge to achieving such goal is the complex nature of human societies which themselves are intertwined within a larger complex environmental system. Socio-environmental loop feedback processes, non-linear relationships among human and environmental phenomena (e.g. climate, culture, economics, politics, etc.), self-organization, among others, are all features of complex systems. In my presentation I will describe a proposal for a complex systems approach for the study of the human past consisting of the following elements:

Big historical Data: achieving high data intensity and variety necessary to characterize a variety of human and natural phenomena. This opens the way to data-driven causal inference.

From proxy to phenomena: proxy data is modelled into the phenomena of interest (e.g. caloric intakes, temperature). Self-developed Bayesian modelling tools are employed to characterize phenomena uncertainty that can be reduced by integrating diverse sources of data.

Data analysis: human and natural phenomena are described in space and time at high resolutions. Machine learning techniques are employed to reconstruct causal webs.